{"id": "c-darwin-01-dessert-01", "uid": "c000", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "prompt_sha256_by_order": ["ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d"], "choice": "sagan", "probs": {"darwin": 0.015708486428544503, "einstein": 0.010747448331157154, "hypatia": 0.005768560479247488, "ibn_sina": 0.00587668534827116, "sagan": 0.9544438238763697, "socrates": 0.007454995536409943}, "confidence": 0.8937383989676638, "seconds": 0.18978872400020919, "seconds_all_orders": 1.1387323440012551, "seconds_by_order": [0.1959931860001234, 0.19066122000003816, 0.18774866400053725, 0.18776486400020076, 0.18775754400121514, 0.18880686599914043], "argmax_stable": true, "spread": {"darwin": 0.006682776365828678, "einstein": 0.008636073389299701, "hypatia": 0.00254438791293452, "ibn_sina": 0.0036951373534431274, "sagan": 0.021304573582594966, "socrates": 0.0044530432736244055}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.020000000000000018, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.015708486428544503, "einstein": 0.010747448331157154, "hypatia": 0.005768560479247488, "ibn_sina": 0.00587668534827116, "sagan": 0.9544438238763697, "socrates": 0.007454995536409943}, {"ibn_sina": 0.006387196330360341, "hypatia": 0.007606144813314453, "sagan": 0.9407873187807674, "socrates": 0.00853799000787879, "darwin": 0.01729782834722222, "einstein": 0.019383521720456855}, {"hypatia": 0.007253374725185089, "darwin": 0.014931270076741377, "sagan": 0.9463959841253942, "ibn_sina": 0.006295328545189131, "einstein": 0.01645662620706017, "socrates": 0.008667416320430072}, {"ibn_sina": 0.007521531268436946, "einstein": 0.017574447378311195, "darwin": 0.01490838145977336, "sagan": 0.9449582736670268, "hypatia": 0.007532034652206541, "socrates": 0.007505331574245164}, {"einstein": 0.014724785347270012, "sagan": 0.9553921878222147, "socrates": 0.0076130121580573725, "ibn_sina": 0.005638789906578384, "darwin": 0.010615051981393543, "hypatia": 0.006016172784486108}, {"socrates": 0.0042143730468056665, "ibn_sina": 0.0038263939149938185, "hypatia": 0.005061756900379933, "einstein": 0.01135174022170417, "sagan": 0.9620918923633623, "darwin": 0.013453843552754051}], "prompt_tokens": 465, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.015708486428544503, "einstein": 0.010747448331157154, "hypatia": 0.005768560479247488, "ibn_sina": 0.00587668534827116, "sagan": 0.9544438238763697, "socrates": 0.007454995536409943}, "confidence": 0.8937383989676638}}, "usage": {"input_tokens": 465, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.006387196330360341, "hypatia": 0.007606144813314453, "sagan": 0.9407873187807674, "socrates": 0.00853799000787879, "darwin": 0.01729782834722222, "einstein": 0.019383521720456855}, "confidence": 0.8631127143373372}}, "usage": {"input_tokens": 465, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.007253374725185089, "darwin": 0.014931270076741377, "sagan": 0.9463959841253942, "ibn_sina": 0.006295328545189131, "einstein": 0.01645662620706017, "socrates": 0.008667416320430072}, "confidence": 0.8755917866235572}}, "usage": {"input_tokens": 465, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.007521531268436946, "einstein": 0.017574447378311195, "darwin": 0.01490838145977336, "sagan": 0.9449582736670268, "hypatia": 0.007532034652206541, "socrates": 0.007505331574245164}, "confidence": 0.872376273989001}}, "usage": {"input_tokens": 465, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.014724785347270012, "sagan": 0.9553921878222147, "socrates": 0.0076130121580573725, "ibn_sina": 0.005638789906578384, "darwin": 0.010615051981393543, "hypatia": 0.006016172784486108}, "confidence": 0.8958756153094131}}, "usage": {"input_tokens": 465, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.0042143730468056665, "ibn_sina": 0.0038263939149938185, "hypatia": 0.005061756900379933, "einstein": 0.01135174022170417, "sagan": 0.9620918923633623, "darwin": 0.013453843552754051}, "confidence": 0.911186441052649}}, "usage": {"input_tokens": 465, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:17Z"}
{"id": "c-darwin-01-dessert-01", "uid": "c001", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "prompt_sha256_by_order": ["ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66"], "choice": "darwin", "probs": {"darwin": 0.9687431215365713, "einstein": 0.01029428200975448, "hypatia": 0.004642326636864988, "ibn_sina": 0.0037227464202851794, "sagan": 0.006762100762565366, "socrates": 0.005835422633958765}, "confidence": 0.9264212751616835, "seconds": 0.22666001683349654, "seconds_all_orders": 1.3599601010009792, "seconds_by_order": [0.22815408099995693, 0.2262543070009997, 0.22610685799918429, 0.22597436000069138, 0.22670035299961455, 0.22677014200053236], "argmax_stable": true, "spread": {"darwin": 0.02082501283492888, "einstein": 0.008188713114230473, "hypatia": 0.0033846315083320457, "ibn_sina": 0.0027176205928342786, "sagan": 0.001658039732646722, "socrates": 0.007334526703367372}, "spread_2dp": {"darwin": 0.020000000000000018, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9687431215365713, "einstein": 0.01029428200975448, "hypatia": 0.004642326636864988, "ibn_sina": 0.0037227464202851794, "sagan": 0.006762100762565366, "socrates": 0.005835422633958765}, {"ibn_sina": 0.00328349064713831, "hypatia": 0.0030421972116719785, "sagan": 0.006028487296216028, "socrates": 0.005867631379091388, "darwin": 0.9730337839898789, "einstein": 0.008744409476003344}, {"hypatia": 0.005972274069878375, "darwin": 0.9622849329517607, "sagan": 0.0071440209093556984, "ibn_sina": 0.004318556257009535, "einstein": 0.01474217296281674, "socrates": 0.005538042849178904}, {"ibn_sina": 0.005412139525353507, "einstein": 0.015648121316875704, "darwin": 0.9558049481320109, "sagan": 0.007161405705005862, "hypatia": 0.0043439999559431244, "socrates": 0.011629385364810877}, {"einstein": 0.0075632114471229, "sagan": 0.006174105390406082, "socrates": 0.004650339973480808, "ibn_sina": 0.0056372299410219024, "darwin": 0.9695482845279644, "hypatia": 0.006426828720004024}, {"socrates": 0.0042948586614435045, "ibn_sina": 0.002919609348187624, "hypatia": 0.003192796848424655, "einstein": 0.007459408202645231, "sagan": 0.00550336597235914, "darwin": 0.9766299609669398}], "prompt_tokens": 657, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.9687431215365713, "einstein": 0.01029428200975448, "hypatia": 0.004642326636864988, "ibn_sina": 0.0037227464202851794, "sagan": 0.006762100762565366, "socrates": 0.005835422633958765}, "confidence": 0.9264212751616835}}, "usage": {"input_tokens": 657, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.00328349064713831, "hypatia": 0.0030421972116719785, "sagan": 0.006028487296216028, "socrates": 0.005867631379091388, "darwin": 0.9730337839898789, "einstein": 0.008744409476003344}, "confidence": 0.9363544210174389}}, "usage": {"input_tokens": 657, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.005972274069878375, "darwin": 0.9622849329517607, "sagan": 0.0071440209093556984, "ibn_sina": 0.004318556257009535, "einstein": 0.01474217296281674, "socrates": 0.005538042849178904}, "confidence": 0.9116147785464441}}, "usage": {"input_tokens": 657, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.005412139525353507, "einstein": 0.015648121316875704, "darwin": 0.9558049481320109, "sagan": 0.007161405705005862, "hypatia": 0.0043439999559431244, "socrates": 0.011629385364810877}, "confidence": 0.8968511829998533}}, "usage": {"input_tokens": 657, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.0075632114471229, "sagan": 0.006174105390406082, "socrates": 0.004650339973480808, "ibn_sina": 0.0056372299410219024, "darwin": 0.9695482845279644, "hypatia": 0.006426828720004024}, "confidence": 0.9282566871118689}}, "usage": {"input_tokens": 657, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.0042948586614435045, "ibn_sina": 0.002919609348187624, "hypatia": 0.003192796848424655, "einstein": 0.007459408202645231, "sagan": 0.00550336597235914, "darwin": 0.9766299609669398}, "confidence": 0.9447150092169145}}, "usage": {"input_tokens": 657, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:18Z"}
{"id": "c-darwin-01-dessert-01", "uid": "c002", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "prompt_sha256_by_order": ["6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc"], "choice": "sagan", "probs": {"darwin": 0.13719805712032074, "einstein": 0.11956542532077158, "hypatia": 0.04345931959038293, "ibn_sina": 0.16953645385221605, "sagan": 0.4687679763748481, "socrates": 0.061472767741460585}, "confidence": 0.1447274035625646, "seconds": 0.18395408849937667, "seconds_all_orders": 1.10372453099626, "seconds_by_order": [0.18431826300002285, 0.18372105699927488, 0.18372270699910587, 0.1839951049987576, 0.18400783399920329, 0.18395956499989552], "argmax_stable": true, "spread": {"darwin": 0.09331810016759717, "einstein": 0.06024002342870957, "hypatia": 0.01773452435815079, "ibn_sina": 0.07190428975445255, "sagan": 0.12995130789215714, "socrates": 0.04218619868098416}, "spread_2dp": {"darwin": 0.09, "einstein": 0.06, "hypatia": 0.010000000000000002, "ibn_sina": 0.07, "sagan": 0.12999999999999995, "socrates": 0.04000000000000001}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.13719805712032074, "einstein": 0.11956542532077158, "hypatia": 0.04345931959038293, "ibn_sina": 0.16953645385221605, "sagan": 0.4687679763748481, "socrates": 0.061472767741460585}, {"ibn_sina": 0.1361605535466024, "hypatia": 0.03410750370433112, "sagan": 0.5667479010538539, "socrates": 0.058670672121338426, "darwin": 0.09919262193322193, "einstein": 0.10512074764065225}, {"hypatia": 0.030495779981038758, "darwin": 0.1336324682649474, "sagan": 0.563822381494324, "ibn_sina": 0.10924913371744532, "einstein": 0.09473866816726903, "socrates": 0.06806156837497564}, {"ibn_sina": 0.0976321640977635, "einstein": 0.11747955165030144, "darwin": 0.1925107221008191, "sagan": 0.4545780234414154, "hypatia": 0.03694266790737803, "socrates": 0.10085687080232258}, {"einstein": 0.10526730736268405, "sagan": 0.5092178541448434, "socrates": 0.0862003099281679, "ibn_sina": 0.14219609369386538, "darwin": 0.13139363963820705, "hypatia": 0.025724795232232145}, {"socrates": 0.07348143265839108, "ibn_sina": 0.1325403171019128, "hypatia": 0.03906217860809495, "einstein": 0.1549786915959786, "sagan": 0.4367965931616967, "darwin": 0.16314078687392575}], "prompt_tokens": 499, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.13719805712032074, "einstein": 0.11956542532077158, "hypatia": 0.04345931959038293, "ibn_sina": 0.16953645385221605, "sagan": 0.4687679763748481, "socrates": 0.061472767741460585}, "confidence": 0.1447274035625646}}, "usage": {"input_tokens": 499, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.1361605535466024, "hypatia": 0.03410750370433112, "sagan": 0.5667479010538539, "socrates": 0.058670672121338426, "darwin": 0.09919262193322193, "einstein": 0.10512074764065225}, "confidence": 0.2382855965172322}}, "usage": {"input_tokens": 499, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.030495779981038758, "darwin": 0.1336324682649474, "sagan": 0.563822381494324, "ibn_sina": 0.10924913371744532, "einstein": 0.09473866816726903, "socrates": 0.06806156837497564}, "confidence": 0.234671767118736}}, "usage": {"input_tokens": 499, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.0976321640977635, "einstein": 0.11747955165030144, "darwin": 0.1925107221008191, "sagan": 0.4545780234414154, "hypatia": 0.03694266790737803, "socrates": 0.10085687080232258}, "confidence": 0.13428629337064296}}, "usage": {"input_tokens": 499, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.10526730736268405, "sagan": 0.5092178541448434, "socrates": 0.0862003099281679, "ibn_sina": 0.14219609369386538, "darwin": 0.13139363963820705, "hypatia": 0.025724795232232145}, "confidence": 0.179152366119869}}, "usage": {"input_tokens": 499, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.07348143265839108, "ibn_sina": 0.1325403171019128, "hypatia": 0.03906217860809495, "einstein": 0.1549786915959786, "sagan": 0.4367965931616967, "darwin": 0.16314078687392575}, "confidence": 0.11910026246671106}}, "usage": {"input_tokens": 499, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:19Z"}
{"id": "c-darwin-01-dessert-02", "uid": "c003", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "prompt_sha256_by_order": ["1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2"], "choice": "darwin", "probs": {"darwin": 0.8156730491631565, "einstein": 0.03359130001571727, "hypatia": 0.019168261521300935, "ibn_sina": 0.016941507912437474, "sagan": 0.09443419488048516, "socrates": 0.020191686506902604}, "confidence": 0.611717028249618, "seconds": 0.2525649685003373, "seconds_all_orders": 1.5153898110020236, "seconds_by_order": [0.2528260680010135, 0.2519288460007374, 0.25273716899937426, 0.25303372599955765, 0.25267314900156634, 0.2521908529997745], "argmax_stable": true, "spread": {"darwin": 0.19078262731570916, "einstein": 0.006757441738492853, "hypatia": 0.009602640558684067, "ibn_sina": 0.0039189100734283305, "sagan": 0.1821808047439047, "socrates": 0.005213004864023485}, "spread_2dp": {"darwin": 0.18999999999999995, "einstein": 0.010000000000000002, "hypatia": 0.009999999999999998, "ibn_sina": 0.0, "sagan": 0.19, "socrates": 0.0}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.8156730491631565, "einstein": 0.03359130001571727, "hypatia": 0.019168261521300935, "ibn_sina": 0.016941507912437474, "sagan": 0.09443419488048516, "socrates": 0.020191686506902604}, {"ibn_sina": 0.016958613739213008, "hypatia": 0.027373588436370402, "sagan": 0.23741695711627864, "socrates": 0.02353894898502297, "darwin": 0.658259260760732, "einstein": 0.03645263096238295}, {"hypatia": 0.0266764127416081, "darwin": 0.7769578539381338, "sagan": 0.12573409818107897, "ibn_sina": 0.017820762616639396, "einstein": 0.03446507305567453, "socrates": 0.018345799466865195}, {"ibn_sina": 0.01974831873545191, "einstein": 0.03412115351289999, "darwin": 0.6333679976827452, "sagan": 0.26555042213640356, "hypatia": 0.023653303601610658, "socrates": 0.02355880433088868}, {"einstein": 0.030329094316177533, "sagan": 0.13068212348601027, "socrates": 0.01892188944661557, "ibn_sina": 0.020860417985865805, "darwin": 0.7704355726853458, "hypatia": 0.028770902079985002}, {"socrates": 0.019929095463809984, "ibn_sina": 0.019759862563917263, "hypatia": 0.023095610357429436, "einstein": 0.029695189223890098, "sagan": 0.08336961739249887, "darwin": 0.8241506249984544}], "prompt_tokens": 701, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.8156730491631565, "einstein": 0.03359130001571727, "hypatia": 0.019168261521300935, "ibn_sina": 0.016941507912437474, "sagan": 0.09443419488048516, "socrates": 0.020191686506902604}, "confidence": 0.611717028249618}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.016958613739213008, "hypatia": 0.027373588436370402, "sagan": 0.23741695711627864, "socrates": 0.02353894898502297, "darwin": 0.658259260760732, "einstein": 0.03645263096238295}, "confidence": 0.39111022030102854}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.0266764127416081, "darwin": 0.7769578539381338, "sagan": 0.12573409818107897, "ibn_sina": 0.017820762616639396, "einstein": 0.03446507305567453, "socrates": 0.018345799466865195}, "confidence": 0.5464314285248444}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.01974831873545191, "einstein": 0.03412115351289999, "darwin": 0.6333679976827452, "sagan": 0.26555042213640356, "hypatia": 0.023653303601610658, "socrates": 0.02355880433088868}, "confidence": 0.3692089509138052}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.030329094316177533, "sagan": 0.13068212348601027, "socrates": 0.01892188944661557, "ibn_sina": 0.020860417985865805, "darwin": 0.7704355726853458, "hypatia": 0.028770902079985002}, "confidence": 0.5358275233181443}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.019929095463809984, "ibn_sina": 0.019759862563917263, "hypatia": 0.023095610357429436, "einstein": 0.029695189223890098, "sagan": 0.08336961739249887, "darwin": 0.8241506249984544}, "confidence": 0.6260530939418679}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:21Z"}
{"id": "c-darwin-01-dessert-03", "uid": "c004", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "prompt_sha256_by_order": ["7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c"], "choice": "sagan", "probs": {"darwin": 0.07395938086089529, "einstein": 0.23396426602491405, "hypatia": 0.01487052280731721, "ibn_sina": 0.015567578057757636, "sagan": 0.6429145310933153, "socrates": 0.01872372115580048}, "confidence": 0.36923490610469045, "seconds": 0.25478059850017115, "seconds_all_orders": 1.5286835910010268, "seconds_by_order": [0.2546714919990336, 0.2550002099997073, 0.25488433099963004, 0.2539978990007512, 0.2551302290012245, 0.25499943000068015], "argmax_stable": false, "spread": {"darwin": 0.05008121758186464, "einstein": 0.3715995034810851, "hypatia": 0.004311020337027176, "ibn_sina": 0.00814301486395436, "sagan": 0.3867070339747091, "socrates": 0.007567463232472427}, "spread_2dp": {"darwin": 0.04999999999999999, "einstein": 0.37, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.38, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "einstein", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.07395938086089529, "einstein": 0.23396426602491405, "hypatia": 0.01487052280731721, "ibn_sina": 0.015567578057757636, "sagan": 0.6429145310933153, "socrates": 0.01872372115580048}, {"ibn_sina": 0.01352350788002762, "hypatia": 0.016870467429234696, "sagan": 0.6288669453245455, "socrates": 0.012957367230158341, "darwin": 0.08838718970816772, "einstein": 0.23939452242786619}, {"hypatia": 0.01255944709220752, "darwin": 0.08947141959309152, "sagan": 0.34787598190198726, "ibn_sina": 0.014297631912772045, "einstein": 0.5183657849348575, "socrates": 0.017429734565084194}, {"ibn_sina": 0.014349648079340298, "einstein": 0.24906685258287414, "darwin": 0.07150855790824638, "sagan": 0.6352347189225351, "hypatia": 0.013381652689115538, "socrates": 0.01645856981788847}, {"einstein": 0.26716734192276725, "sagan": 0.5774852824834474, "socrates": 0.011156257923328052, "ibn_sina": 0.009455281760370558, "darwin": 0.1201059752317498, "hypatia": 0.014629860678336918}, {"socrates": 0.015365972377235092, "ibn_sina": 0.017598296624324918, "hypatia": 0.01566167601808603, "einstein": 0.1467662814537724, "sagan": 0.7345830158766964, "darwin": 0.07002475764988517}], "prompt_tokens": 721, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.07395938086089529, "einstein": 0.23396426602491405, "hypatia": 0.01487052280731721, "ibn_sina": 0.015567578057757636, "sagan": 0.6429145310933153, "socrates": 0.01872372115580048}, "confidence": 0.36923490610469045}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.01352350788002762, "hypatia": 0.016870467429234696, "sagan": 0.6288669453245455, "socrates": 0.012957367230158341, "darwin": 0.08838718970816772, "einstein": 0.23939452242786619}, "confidence": 0.3534772706763974}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.01255944709220752, "darwin": 0.08947141959309152, "sagan": 0.34787598190198726, "ibn_sina": 0.014297631912772045, "einstein": 0.5183657849348575, "socrates": 0.017429734565084194}, "confidence": 0.27807025400318813}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.014349648079340298, "einstein": 0.24906685258287414, "darwin": 0.07150855790824638, "sagan": 0.6352347189225351, "hypatia": 0.013381652689115538, "socrates": 0.01645856981788847}, "confidence": 0.3655921415000701}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.26716734192276725, "sagan": 0.5774852824834474, "socrates": 0.011156257923328052, "ibn_sina": 0.009455281760370558, "darwin": 0.1201059752317498, "hypatia": 0.014629860678336918}, "confidence": 0.3036651791547052}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.015365972377235092, "ibn_sina": 0.017598296624324918, "hypatia": 0.01566167601808603, "einstein": 0.1467662814537724, "sagan": 0.7345830158766964, "darwin": 0.07002475764988517}, "confidence": 0.4802165398204804}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:22Z"}
{"id": "c-darwin-01-dessert-03", "uid": "c005", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "prompt_sha256_by_order": ["1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06"], "choice": "darwin", "probs": {"darwin": 0.9763350331289895, "einstein": 0.004977643578653962, "hypatia": 0.0047689423422601, "ibn_sina": 0.0024123381119557584, "sagan": 0.0061345008756890085, "socrates": 0.005371541962451732}, "confidence": 0.9440199059210315, "seconds": 0.2549782188334575, "seconds_all_orders": 1.529869313000745, "seconds_by_order": [0.2549507410003571, 0.2541149280004902, 0.25577528400026495, 0.2557220939997933, 0.25506954999946174, 0.2542367160003778], "argmax_stable": true, "spread": {"darwin": 0.009446613801227133, "einstein": 0.0025982265946971755, "hypatia": 0.003624467349577696, "ibn_sina": 0.0021955997362500183, "sagan": 0.0025299100014015297, "socrates": 0.005563197976327203}, "spread_2dp": {"darwin": 0.010000000000000009, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9763350331289895, "einstein": 0.004977643578653962, "hypatia": 0.0047689423422601, "ibn_sina": 0.0024123381119557584, "sagan": 0.0061345008756890085, "socrates": 0.005371541962451732}, {"ibn_sina": 0.002592175061541425, "hypatia": 0.0034759572906877863, "sagan": 0.0041390738242552735, "socrates": 0.0047199495804612305, "darwin": 0.9790030565140688, "einstein": 0.006069787728985391}, {"hypatia": 0.00647931299569637, "darwin": 0.9757263840262307, "sagan": 0.006286876560020364, "ibn_sina": 0.003013761363146189, "einstein": 0.003985059832093493, "socrates": 0.004508605222812838}, {"ibn_sina": 0.004183461817219198, "einstein": 0.005819202091334737, "darwin": 0.9695564427128417, "sagan": 0.0056439104911483115, "hypatia": 0.0053845497537510405, "socrates": 0.009412433133705043}, {"einstein": 0.0034715611342882157, "sagan": 0.004682339289988675, "socrates": 0.0038492351573778403, "ibn_sina": 0.004607937848205777, "darwin": 0.9765951668127979, "hypatia": 0.0067937597573415245}, {"socrates": 0.0049298079173633845, "ibn_sina": 0.00304571207013939, "hypatia": 0.0031692924077638285, "einstein": 0.004109285995187509, "sagan": 0.006668983825656803, "darwin": 0.978076917783889}], "prompt_tokens": 725, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.9763350331289895, "einstein": 0.004977643578653962, "hypatia": 0.0047689423422601, "ibn_sina": 0.0024123381119557584, "sagan": 0.0061345008756890085, "socrates": 0.005371541962451732}, "confidence": 0.9440199059210315}}, "usage": {"input_tokens": 725, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.002592175061541425, "hypatia": 0.0034759572906877863, "sagan": 0.0041390738242552735, "socrates": 0.0047199495804612305, "darwin": 0.9790030565140688, "einstein": 0.006069787728985391}, "confidence": 0.9502504461925181}}, "usage": {"input_tokens": 725, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.00647931299569637, "darwin": 0.9757263840262307, "sagan": 0.006286876560020364, "ibn_sina": 0.003013761363146189, "einstein": 0.003985059832093493, "socrates": 0.004508605222812838}, "confidence": 0.9426025285349862}}, "usage": {"input_tokens": 725, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.004183461817219198, "einstein": 0.005819202091334737, "darwin": 0.9695564427128417, "sagan": 0.0056439104911483115, "hypatia": 0.0053845497537510405, "socrates": 0.009412433133705043}, "confidence": 0.928288601285098}}, "usage": {"input_tokens": 725, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.0034715611342882157, "sagan": 0.004682339289988675, "socrates": 0.0038492351573778403, "ibn_sina": 0.004607937848205777, "darwin": 0.9765951668127979, "hypatia": 0.0067937597573415245}, "confidence": 0.9446251609050428}}, "usage": {"input_tokens": 725, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.0049298079173633845, "ibn_sina": 0.00304571207013939, "hypatia": 0.0031692924077638285, "einstein": 0.004109285995187509, "sagan": 0.006668983825656803, "darwin": 0.978076917783889}, "confidence": 0.948087330952926}}, "usage": {"input_tokens": 725, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:24Z"}
{"id": "c-darwin-01-main-01", "uid": "c006", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "prompt_sha256_by_order": ["56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545"], "choice": "ibn_sina", "probs": {"darwin": 0.3619844869867786, "einstein": 0.018691862824938153, "hypatia": 0.014974670292635633, "ibn_sina": 0.5848671012118115, "sagan": 0.01117509620581287, "socrates": 0.008306782478023238}, "confidence": 0.3686437681546894, "seconds": 0.1881220873331889, "seconds_all_orders": 1.1287325239991333, "seconds_by_order": [0.18870851599967864, 0.18797500199980277, 0.1880880310000066, 0.18816847999914899, 0.1874767259996588, 0.18831576900083746], "argmax_stable": false, "spread": {"darwin": 0.38342634350343524, "einstein": 0.01094671326155393, "hypatia": 0.009062793457010354, "ibn_sina": 0.3748458902263032, "sagan": 0.006646839392504735, "socrates": 0.005144136010595323}, "spread_2dp": {"darwin": 0.39, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.37, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["ibn_sina", "darwin", "darwin", "darwin", "darwin", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.3619844869867786, "einstein": 0.018691862824938153, "hypatia": 0.014974670292635633, "ibn_sina": 0.5848671012118115, "sagan": 0.01117509620581287, "socrates": 0.008306782478023238}, {"ibn_sina": 0.3832618320068025, "hypatia": 0.010750356656400385, "sagan": 0.00598445747831333, "socrates": 0.004920255713750156, "darwin": 0.5865722843394178, "einstein": 0.008510813805315931}, {"hypatia": 0.018803491891075344, "darwin": 0.5117232848975862, "sagan": 0.008454902222510842, "ibn_sina": 0.44862427731706617, "einstein": 0.007745149563384223, "socrates": 0.0046488941083771855}, {"ibn_sina": 0.4719765474206149, "einstein": 0.01339267182011962, "darwin": 0.48650025220161797, "sagan": 0.0064115464839976126, "hypatia": 0.01192595195467746, "socrates": 0.009793030118972508}, {"einstein": 0.012404937146899277, "sagan": 0.0067444358583907595, "socrates": 0.00572723351395352, "ibn_sina": 0.2100212109855083, "darwin": 0.7454108304902138, "hypatia": 0.019691352005034267}, {"socrates": 0.008044032734864462, "ibn_sina": 0.5344920387985042, "hypatia": 0.010628558548023912, "einstein": 0.018140308505612886, "sagan": 0.012631296870818065, "darwin": 0.4160637645421765}], "prompt_tokens": 535, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.3619844869867786, "einstein": 0.018691862824938153, "hypatia": 0.014974670292635633, "ibn_sina": 0.5848671012118115, "sagan": 0.01117509620581287, "socrates": 0.008306782478023238}, "confidence": 0.3686437681546894}}, "usage": {"input_tokens": 535, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.3832618320068025, "hypatia": 0.010750356656400385, "sagan": 0.00598445747831333, "socrates": 0.004920255713750156, "darwin": 0.5865722843394178, "einstein": 0.008510813805315931}, "confidence": 0.38944564407500837}}, "usage": {"input_tokens": 535, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.018803491891075344, "darwin": 0.5117232848975862, "sagan": 0.008454902222510842, "ibn_sina": 0.44862427731706617, "einstein": 0.007745149563384223, "socrates": 0.0046488941083771855}, "confidence": 0.35635734249005935}}, "usage": {"input_tokens": 535, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.4719765474206149, "einstein": 0.01339267182011962, "darwin": 0.48650025220161797, "sagan": 0.0064115464839976126, "hypatia": 0.01192595195467746, "socrates": 0.009793030118972508}, "confidence": 0.35188355207566724}}, "usage": {"input_tokens": 535, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.012404937146899277, "sagan": 0.0067444358583907595, "socrates": 0.00572723351395352, "ibn_sina": 0.2100212109855083, "darwin": 0.7454108304902138, "hypatia": 0.019691352005034267}, "confidence": 0.5204393628449269}}, "usage": {"input_tokens": 535, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.008044032734864462, "ibn_sina": 0.5344920387985042, "hypatia": 0.010628558548023912, "einstein": 0.018140308505612886, "sagan": 0.012631296870818065, "darwin": 0.4160637645421765}, "confidence": 0.35134850665210626}}, "usage": {"input_tokens": 535, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:25Z"}
{"id": "c-darwin-01-main-02", "uid": "c007", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "prompt_sha256_by_order": ["7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf"], "choice": "sagan", "probs": {"darwin": 0.1405351795508707, "einstein": 0.08000728946754945, "hypatia": 0.0123279249874294, "ibn_sina": 0.02196785586944515, "sagan": 0.721709610904065, "socrates": 0.023452139220640357}, "confidence": 0.45784075934878893, "seconds": 0.1837734051659936, "seconds_all_orders": 1.1026404309959617, "seconds_by_order": [0.18398072499985574, 0.18370546699952683, 0.1839172359996155, 0.1830902329984383, 0.1838944459996128, 0.18405232399891247], "argmax_stable": false, "spread": {"darwin": 0.34749295426597515, "einstein": 0.04974169890496408, "hypatia": 0.0034604218144188674, "ibn_sina": 0.012549453431950266, "sagan": 0.436088378290991, "socrates": 0.05164000722282607}, "spread_2dp": {"darwin": 0.35000000000000003, "einstein": 0.05, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.44, "socrates": 0.05}, "choices_by_order": ["sagan", "sagan", "darwin", "darwin", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.1405351795508707, "einstein": 0.08000728946754945, "hypatia": 0.0123279249874294, "ibn_sina": 0.02196785586944515, "sagan": 0.721709610904065, "socrates": 0.023452139220640357}, {"ibn_sina": 0.010915586875790908, "hypatia": 0.008867503173010532, "sagan": 0.7083433312758031, "socrates": 0.036855142280846215, "darwin": 0.1571063903957839, "einstein": 0.07791204599876538}, {"hypatia": 0.010562464895090865, "darwin": 0.4611662303043074, "sagan": 0.39889140208572105, "ibn_sina": 0.014547056827262716, "einstein": 0.09007613423206748, "socrates": 0.02475671165555053}, {"ibn_sina": 0.018605616722666887, "einstein": 0.10347558034294145, "darwin": 0.4486603946416564, "sagan": 0.34341148539833255, "hypatia": 0.011942721324810963, "socrates": 0.07390420156959168}, {"einstein": 0.053733881437977374, "sagan": 0.7794998636893236, "socrates": 0.03413870983401133, "ibn_sina": 0.009418402437494883, "darwin": 0.1136732760383323, "hypatia": 0.009535866562860491}, {"socrates": 0.02226419434676561, "ibn_sina": 0.013837964250385793, "hypatia": 0.010347470716169729, "einstein": 0.08861090936692866, "sagan": 0.617905074045479, "darwin": 0.2470343872742712}], "prompt_tokens": 497, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.1405351795508707, "einstein": 0.08000728946754945, "hypatia": 0.0123279249874294, "ibn_sina": 0.02196785586944515, "sagan": 0.721709610904065, "socrates": 0.023452139220640357}, "confidence": 0.45784075934878893}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.010915586875790908, "hypatia": 0.008867503173010532, "sagan": 0.7083433312758031, "socrates": 0.036855142280846215, "darwin": 0.1571063903957839, "einstein": 0.07791204599876538}, "confidence": 0.44087087672765923}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.010562464895090865, "darwin": 0.4611662303043074, "sagan": 0.39889140208572105, "ibn_sina": 0.014547056827262716, "einstein": 0.09007613423206748, "socrates": 0.02475671165555053}, "confidence": 0.2570061158658197}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.018605616722666887, "einstein": 0.10347558034294145, "darwin": 0.4486603946416564, "sagan": 0.34341148539833255, "hypatia": 0.011942721324810963, "socrates": 0.07390420156959168}, "confidence": 0.2030625067920596}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.053733881437977374, "sagan": 0.7794998636893236, "socrates": 0.03413870983401133, "ibn_sina": 0.009418402437494883, "darwin": 0.1136732760383323, "hypatia": 0.009535866562860491}, "confidence": 0.549728886107274}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.02226419434676561, "ibn_sina": 0.013837964250385793, "hypatia": 0.010347470716169729, "einstein": 0.08861090936692866, "sagan": 0.617905074045479, "darwin": 0.2470343872742712}, "confidence": 0.3417745792488221}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:26Z"}
{"id": "c-darwin-01-main-02", "uid": "c008", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "prompt_sha256_by_order": ["99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce"], "choice": "darwin", "probs": {"darwin": 0.8998516969935327, "einstein": 0.004574592056540266, "hypatia": 0.005901874679061458, "ibn_sina": 0.003337879218695834, "sagan": 0.005850446513952388, "socrates": 0.0804835105382173}, "confidence": 0.7795741602759978, "seconds": 0.22567772966673752, "seconds_all_orders": 1.3540663780004252, "seconds_by_order": [0.225490152999555, 0.22602220800035866, 0.225226405000285, 0.22590101999958279, 0.2257064410005114, 0.22572015100013232], "argmax_stable": true, "spread": {"darwin": 0.10411994727713081, "einstein": 0.0038275199089739986, "hypatia": 0.00726450584892626, "ibn_sina": 0.002681953956289352, "sagan": 0.0013857439074222875, "socrates": 0.09336668624584239}, "spread_2dp": {"darwin": 0.09999999999999998, "einstein": 0.01, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.01, "socrates": 0.09}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.8998516969935327, "einstein": 0.004574592056540266, "hypatia": 0.005901874679061458, "ibn_sina": 0.003337879218695834, "sagan": 0.005850446513952388, "socrates": 0.0804835105382173}, {"ibn_sina": 0.005389213532526302, "hypatia": 0.011003849798329132, "sagan": 0.004487001155142088, "socrates": 0.04921893984789298, "darwin": 0.9227509940308662, "einstein": 0.0071500016352432335}, {"hypatia": 0.011012983438538106, "darwin": 0.9203249661993449, "sagan": 0.0044647026065301005, "ibn_sina": 0.0032778591352002786, "einstein": 0.0045595188572981876, "socrates": 0.05635996976308843}, {"ibn_sina": 0.0059598130914896305, "einstein": 0.0049789092287717, "darwin": 0.8358612475596104, "sagan": 0.00534272239291015, "hypatia": 0.012956971546486208, "socrates": 0.13490033618073188}, {"einstein": 0.004038036810779508, "sagan": 0.005204377846391232, "socrates": 0.06510217275052683, "ibn_sina": 0.005426091475870366, "darwin": 0.9079957840741038, "hypatia": 0.012233537042328239}, {"socrates": 0.04153364993488949, "ibn_sina": 0.004188762734484085, "hypatia": 0.005692465697559948, "einstein": 0.003322481726269235, "sagan": 0.005281445070055931, "darwin": 0.9399811948367413}], "prompt_tokens": 649, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.8998516969935327, "einstein": 0.004574592056540266, "hypatia": 0.005901874679061458, "ibn_sina": 0.003337879218695834, "sagan": 0.005850446513952388, "socrates": 0.0804835105382173}, "confidence": 0.7795741602759978}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.005389213532526302, "hypatia": 0.011003849798329132, "sagan": 0.004487001155142088, "socrates": 0.04921893984789298, "darwin": 0.9227509940308662, "einstein": 0.0071500016352432335}, "confidence": 0.824935942072403}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.011012983438538106, "darwin": 0.9203249661993449, "sagan": 0.0044647026065301005, "ibn_sina": 0.0032778591352002786, "einstein": 0.0045595188572981876, "socrates": 0.05635996976308843}, "confidence": 0.8204166910573889}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.0059598130914896305, "einstein": 0.0049789092287717, "darwin": 0.8358612475596104, "sagan": 0.00534272239291015, "hypatia": 0.012956971546486208, "socrates": 0.13490033618073188}, "confidence": 0.660542635092401}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.004038036810779508, "sagan": 0.005204377846391232, "socrates": 0.06510217275052683, "ibn_sina": 0.005426091475870366, "darwin": 0.9079957840741038, "hypatia": 0.012233537042328239}, "confidence": 0.7947005559766117}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.04153364993488949, "ibn_sina": 0.004188762734484085, "hypatia": 0.005692465697559948, "einstein": 0.003322481726269235, "sagan": 0.005281445070055931, "darwin": 0.9399811948367413}, "confidence": 0.8624542878033218}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:27Z"}
{"id": "c-darwin-01-main-02", "uid": "c009", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "prompt_sha256_by_order": ["e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1"], "choice": "darwin", "probs": {"darwin": 0.9472631319538175, "einstein": 0.014353853412535127, "hypatia": 0.0077573686915105005, "ibn_sina": 0.005764966688518276, "sagan": 0.012331317677712571, "socrates": 0.012529361575906076}, "confidence": 0.877499118608618, "seconds": 0.2543242073334113, "seconds_all_orders": 1.525945244000468, "seconds_by_order": [0.2542788760001713, 0.25426691599932383, 0.2546149930003594, 0.2544080749994464, 0.2538421700010076, 0.2545342140001594], "argmax_stable": true, "spread": {"darwin": 0.01907481083539886, "einstein": 0.010463326312184138, "hypatia": 0.005196813660936113, "ibn_sina": 0.006551450789326251, "sagan": 0.002374982410998226, "socrates": 0.007848493021138995}, "spread_2dp": {"darwin": 0.010000000000000009, "einstein": 0.01, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9472631319538175, "einstein": 0.014353853412535127, "hypatia": 0.0077573686915105005, "ibn_sina": 0.005764966688518276, "sagan": 0.012331317677712571, "socrates": 0.012529361575906076}, {"ibn_sina": 0.008200499929376348, "hypatia": 0.007029766227756118, "sagan": 0.010412291129562521, "socrates": 0.009058022193609187, "darwin": 0.9454359170205946, "einstein": 0.01986350349910116}, {"hypatia": 0.01012267174331679, "darwin": 0.945481627373522, "sagan": 0.011138118500482797, "ibn_sina": 0.006718748017900662, "einstein": 0.015987605337839553, "socrates": 0.010551229026938197}, {"ibn_sina": 0.012316417477844527, "einstein": 0.016885515785037664, "darwin": 0.9358639265005657, "sagan": 0.010719233713515384, "hypatia": 0.007691968668856949, "socrates": 0.016522937854179777}, {"einstein": 0.012744252272839578, "sagan": 0.012787273540560747, "socrates": 0.008674444833040782, "ibn_sina": 0.007765328248109184, "darwin": 0.9476934776545309, "hypatia": 0.010335223450918708}, {"socrates": 0.011747782723182932, "ibn_sina": 0.007319279157220634, "hypatia": 0.005138409789982595, "einstein": 0.009400177186917023, "sagan": 0.01145561380673216, "darwin": 0.9549387373359646}], "prompt_tokens": 717, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.9472631319538175, "einstein": 0.014353853412535127, "hypatia": 0.0077573686915105005, "ibn_sina": 0.005764966688518276, "sagan": 0.012331317677712571, "socrates": 0.012529361575906076}, "confidence": 0.877499118608618}}, "usage": {"input_tokens": 717, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.008200499929376348, "hypatia": 0.007029766227756118, "sagan": 0.010412291129562521, "socrates": 0.009058022193609187, "darwin": 0.9454359170205946, "einstein": 0.01986350349910116}, "confidence": 0.8734609136185201}}, "usage": {"input_tokens": 717, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.01012267174331679, "darwin": 0.945481627373522, "sagan": 0.011138118500482797, "ibn_sina": 0.006718748017900662, "einstein": 0.015987605337839553, "socrates": 0.010551229026938197}, "confidence": 0.8734889288814195}}, "usage": {"input_tokens": 717, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.012316417477844527, "einstein": 0.016885515785037664, "darwin": 0.9358639265005657, "sagan": 0.010719233713515384, "hypatia": 0.007691968668856949, "socrates": 0.016522937854179777}, "confidence": 0.8520702154439348}}, "usage": {"input_tokens": 717, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.012744252272839578, "sagan": 0.012787273540560747, "socrates": 0.008674444833040782, "ibn_sina": 0.007765328248109184, "darwin": 0.9476934776545309, "hypatia": 0.010335223450918708}, "confidence": 0.8784294652951321}}, "usage": {"input_tokens": 717, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.011747782723182932, "ibn_sina": 0.007319279157220634, "hypatia": 0.005138409789982595, "einstein": 0.009400177186917023, "sagan": 0.01145561380673216, "darwin": 0.9549387373359646}, "confidence": 0.8948146863821236}}, "usage": {"input_tokens": 717, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:29Z"}
{"id": "c-darwin-01-main-02", "uid": "c010", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "prompt_sha256_by_order": ["b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285"], "choice": "darwin", "probs": {"darwin": 0.45009935469820134, "einstein": 0.04084740279187615, "hypatia": 0.04879498650299553, "ibn_sina": 0.43512960749149554, "sagan": 0.006412591116953163, "socrates": 0.018716057398478314}, "confidence": 0.2756416930809831, "seconds": 0.252864049333463, "seconds_all_orders": 1.517184296000778, "seconds_by_order": [0.25301173700063373, 0.25307643599990115, 0.2522275730007095, 0.2532051449998107, 0.252812577999066, 0.25285082700065686], "argmax_stable": false, "spread": {"darwin": 0.23289855056890763, "einstein": 0.03894425640321994, "hypatia": 0.0963371987327781, "ibn_sina": 0.18813346356044458, "sagan": 0.0023174008208079536, "socrates": 0.026920132638029004}, "spread_2dp": {"darwin": 0.24, "einstein": 0.04, "hypatia": 0.1, "ibn_sina": 0.18, "sagan": 0.0, "socrates": 0.019999999999999997}, "choices_by_order": ["darwin", "ibn_sina", "ibn_sina", "ibn_sina", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.45009935469820134, "einstein": 0.04084740279187615, "hypatia": 0.04879498650299553, "ibn_sina": 0.43512960749149554, "sagan": 0.006412591116953163, "socrates": 0.018716057398478314}, {"ibn_sina": 0.5240073978221681, "hypatia": 0.059624228242626136, "sagan": 0.00635785006374316, "socrates": 0.009361304157437397, "darwin": 0.37152559370487914, "einstein": 0.029123626009145996}, {"hypatia": 0.08698888381405159, "darwin": 0.3540905689654148, "sagan": 0.008675250884551114, "ibn_sina": 0.4824474479870638, "einstein": 0.05313479028228901, "socrates": 0.014663058066629701}, {"ibn_sina": 0.4796812182892578, "einstein": 0.02918297702121086, "darwin": 0.3959943937338755, "sagan": 0.007649196244034051, "hypatia": 0.052672108219324325, "socrates": 0.03482010649229751}, {"einstein": 0.04676076944441005, "sagan": 0.006568562171470583, "socrates": 0.00789997385426851, "ibn_sina": 0.35877445473044184, "darwin": 0.4491785397477239, "hypatia": 0.13081770005168503}, {"socrates": 0.02097439626443108, "ibn_sina": 0.3358739342617235, "hypatia": 0.03448050131890693, "einstein": 0.01419053387906907, "sagan": 0.007491514741547058, "darwin": 0.5869891195343224}], "prompt_tokens": 703, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.45009935469820134, "einstein": 0.04084740279187615, "hypatia": 0.04879498650299553, "ibn_sina": 0.43512960749149554, "sagan": 0.006412591116953163, "socrates": 0.018716057398478314}, "confidence": 0.2756416930809831}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.5240073978221681, "hypatia": 0.059624228242626136, "sagan": 0.00635785006374316, "socrates": 0.009361304157437397, "darwin": 0.37152559370487914, "einstein": 0.029123626009145996}, "confidence": 0.30057557225024}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.08698888381405159, "darwin": 0.3540905689654148, "sagan": 0.008675250884551114, "ibn_sina": 0.4824474479870638, "einstein": 0.05313479028228901, "socrates": 0.014663058066629701}, "confidence": 0.24257956983377743}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.4796812182892578, "einstein": 0.02918297702121086, "darwin": 0.3959943937338755, "sagan": 0.007649196244034051, "hypatia": 0.052672108219324325, "socrates": 0.03482010649229751}, "confidence": 0.2701630938394356}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.04676076944441005, "sagan": 0.006568562171470583, "socrates": 0.00789997385426851, "ibn_sina": 0.35877445473044184, "darwin": 0.4491785397477239, "hypatia": 0.13081770005168503}, "confidence": 0.21986303888652153}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.02097439626443108, "ibn_sina": 0.3358739342617235, "hypatia": 0.03448050131890693, "einstein": 0.01419053387906907, "sagan": 0.007491514741547058, "darwin": 0.5869891195343224}, "confidence": 0.35110462057946074}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:30Z"}
{"id": "c-darwin-01-main-03", "uid": "c011", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "prompt_sha256_by_order": ["c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5"], "choice": "darwin", "probs": {"darwin": 0.9675082575159987, "einstein": 0.007465259935239031, "hypatia": 0.005443138716080264, "ibn_sina": 0.0036816214159942615, "sagan": 0.008987626221832131, "socrates": 0.006914096194855605}, "confidence": 0.9235596672570076, "seconds": 0.2535761970002568, "seconds_all_orders": 1.5214571820015408, "seconds_by_order": [0.253039846998945, 0.25411282800087065, 0.2535995220005134, 0.2534691330001806, 0.25316464500065194, 0.2540712070003792], "argmax_stable": true, "spread": {"darwin": 0.015837831851568063, "einstein": 0.003989796211876673, "hypatia": 0.003909550309936113, "ibn_sina": 0.003198073863426339, "sagan": 0.003376678321150171, "socrates": 0.007300154643688377}, "spread_2dp": {"darwin": 0.010000000000000009, "einstein": 0.0, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9675082575159987, "einstein": 0.007465259935239031, "hypatia": 0.005443138716080264, "ibn_sina": 0.0036816214159942615, "sagan": 0.008987626221832131, "socrates": 0.006914096194855605}, {"ibn_sina": 0.004394161340649177, "hypatia": 0.005156614470072389, "sagan": 0.00561094790068196, "socrates": 0.00622999545449316, "darwin": 0.9704150590993481, "einstein": 0.008193221734755327}, {"hypatia": 0.006542357859719811, "darwin": 0.971655798544786, "sagan": 0.007023923856665473, "ibn_sina": 0.0035674627592364797, "einstein": 0.006024664859556085, "socrates": 0.005185792120036172}, {"ibn_sina": 0.006765536622662819, "einstein": 0.010014461071432758, "darwin": 0.955817966693218, "sagan": 0.008293519073985341, "hypatia": 0.0073743252815485485, "socrates": 0.011734191257152726}, {"einstein": 0.008169782518964275, "sagan": 0.007443044271291301, "socrates": 0.004434036613464349, "ibn_sina": 0.00589855824753094, "darwin": 0.9670122348958388, "hypatia": 0.007042343452910355}, {"socrates": 0.006486702143100643, "ibn_sina": 0.0056808888310130815, "hypatia": 0.0034647749716124362, "einstein": 0.006169867649292171, "sagan": 0.007747409136347964, "darwin": 0.9704503572686338}], "prompt_tokens": 707, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.9675082575159987, "einstein": 0.007465259935239031, "hypatia": 0.005443138716080264, "ibn_sina": 0.0036816214159942615, "sagan": 0.008987626221832131, "socrates": 0.006914096194855605}, "confidence": 0.9235596672570076}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.004394161340649177, "hypatia": 0.005156614470072389, "sagan": 0.00561094790068196, "socrates": 0.00622999545449316, "darwin": 0.9704150590993481, "einstein": 0.008193221734755327}, "confidence": 0.9302664528586886}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.006542357859719811, "darwin": 0.971655798544786, "sagan": 0.007023923856665473, "ibn_sina": 0.0035674627592364797, "einstein": 0.006024664859556085, "socrates": 0.005185792120036172}, "confidence": 0.9331396535386425}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.006765536622662819, "einstein": 0.010014461071432758, "darwin": 0.955817966693218, "sagan": 0.008293519073985341, "hypatia": 0.0073743252815485485, "socrates": 0.011734191257152726}, "confidence": 0.8967938820956192}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.008169782518964275, "sagan": 0.007443044271291301, "socrates": 0.004434036613464349, "ibn_sina": 0.00589855824753094, "darwin": 0.9670122348958388, "hypatia": 0.007042343452910355}, "confidence": 0.9224066259568889}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.006486702143100643, "ibn_sina": 0.0056808888310130815, "hypatia": 0.0034647749716124362, "einstein": 0.006169867649292171, "sagan": 0.007747409136347964, "darwin": 0.9704503572686338}, "confidence": 0.9303500080073948}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:32Z"}
{"id": "c-darwin-01-starter-01", "uid": "c012", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "prompt_sha256_by_order": ["97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1"], "choice": "sagan", "probs": {"darwin": 0.021692691218826263, "einstein": 0.1578406546168662, "hypatia": 0.011121814861307302, "ibn_sina": 0.029734381846245148, "sagan": 0.7684317788186048, "socrates": 0.011178678638150353}, "confidence": 0.5404053218634447, "seconds": 0.14741045616635043, "seconds_all_orders": 0.8844627369981026, "seconds_by_order": [0.14758443499886198, 0.1471636980004405, 0.14765959400028805, 0.14741417599907436, 0.14728237700001046, 0.1473584569994273], "argmax_stable": true, "spread": {"darwin": 0.010806767412156942, "einstein": 0.10623569377010278, "hypatia": 0.004279533063612092, "ibn_sina": 0.016086415858337512, "sagan": 0.12076326745173571, "socrates": 0.002640184517860033}, "spread_2dp": {"darwin": 0.01, "einstein": 0.11, "hypatia": 0.0, "ibn_sina": 0.019999999999999997, "sagan": 0.12, "socrates": 0.0}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.021692691218826263, "einstein": 0.1578406546168662, "hypatia": 0.011121814861307302, "ibn_sina": 0.029734381846245148, "sagan": 0.7684317788186048, "socrates": 0.011178678638150353}, {"ibn_sina": 0.013647965987907636, "hypatia": 0.010530268271417075, "sagan": 0.8879544870010916, "socrates": 0.012254522433934717, "darwin": 0.014659658549377008, "einstein": 0.06095309775627198}, {"hypatia": 0.009707872197233732, "darwin": 0.013373886573783939, "sagan": 0.8468626328235034, "ibn_sina": 0.02625113148720922, "einstein": 0.090727589317085, "socrates": 0.01307688760118472}, {"ibn_sina": 0.027900413694848282, "einstein": 0.09346215197515972, "darwin": 0.02418065398594088, "sagan": 0.8312059254116075, "hypatia": 0.011803134074520237, "socrates": 0.011447720857923258}, {"einstein": 0.051604960846763415, "sagan": 0.8891950462703405, "socrates": 0.013632203947833264, "ibn_sina": 0.02043611705239682, "darwin": 0.01587321577745357, "hypatia": 0.009258456105212408}, {"socrates": 0.013818863156010386, "ibn_sina": 0.024762527530301816, "hypatia": 0.0135379891688245, "einstein": 0.14661598080935956, "sagan": 0.781655622453971, "darwin": 0.019609016881532715}], "prompt_tokens": 387, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.021692691218826263, "einstein": 0.1578406546168662, "hypatia": 0.011121814861307302, "ibn_sina": 0.029734381846245148, "sagan": 0.7684317788186048, "socrates": 0.011178678638150353}, "confidence": 0.5404053218634447}}, "usage": {"input_tokens": 387, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.013647965987907636, "hypatia": 0.010530268271417075, "sagan": 0.8879544870010916, "socrates": 0.012254522433934717, "darwin": 0.014659658549377008, "einstein": 0.06095309775627198}, "confidence": 0.7514088202549608}}, "usage": {"input_tokens": 387, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.009707872197233732, "darwin": 0.013373886573783939, "sagan": 0.8468626328235034, "ibn_sina": 0.02625113148720922, "einstein": 0.090727589317085, "socrates": 0.01307688760118472}, "confidence": 0.6718492538253095}}, "usage": {"input_tokens": 387, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.027900413694848282, "einstein": 0.09346215197515972, "darwin": 0.02418065398594088, "sagan": 0.8312059254116075, "hypatia": 0.011803134074520237, "socrates": 0.011447720857923258}, "confidence": 0.6415263588274579}}, "usage": {"input_tokens": 387, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.051604960846763415, "sagan": 0.8891950462703405, "socrates": 0.013632203947833264, "ibn_sina": 0.02043611705239682, "darwin": 0.01587321577745357, "hypatia": 0.009258456105212408}, "confidence": 0.75312646257872}}, "usage": {"input_tokens": 387, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.013818863156010386, "ibn_sina": 0.024762527530301816, "hypatia": 0.0135379891688245, "einstein": 0.14661598080935956, "sagan": 0.781655622453971, "darwin": 0.019609016881532715}, "confidence": 0.5606244308620051}}, "usage": {"input_tokens": 387, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:33Z"}
{"id": "c-darwin-01-starter-01", "uid": "c013", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "prompt_sha256_by_order": ["b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511"], "choice": "darwin", "probs": {"darwin": 0.977843288450736, "einstein": 0.006797375120892521, "hypatia": 0.0033080956404960573, "ibn_sina": 0.0028704981224578546, "sagan": 0.004687513485026962, "socrates": 0.004493229180390756}, "confidence": 0.947542055468873, "seconds": 0.18686098949971588, "seconds_all_orders": 1.1211659369982954, "seconds_by_order": [0.18708714899912593, 0.18673581199982436, 0.18655485399904137, 0.18678991200067685, 0.18703394899966952, 0.18696426099995733], "argmax_stable": true, "spread": {"darwin": 0.019500105388514433, "einstein": 0.010257914137552867, "hypatia": 0.003441667586405756, "ibn_sina": 0.003127103653794037, "sagan": 0.0016175789913586485, "socrates": 0.003658667351419372}, "spread_2dp": {"darwin": 0.020000000000000018, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.977843288450736, "einstein": 0.006797375120892521, "hypatia": 0.0033080956404960573, "ibn_sina": 0.0028704981224578546, "sagan": 0.004687513485026962, "socrates": 0.004493229180390756}, {"ibn_sina": 0.0030862638873911975, "hypatia": 0.002630196393873711, "sagan": 0.004743679755755736, "socrates": 0.004305065428196521, "darwin": 0.9727325689621946, "einstein": 0.012502225572588118}, {"hypatia": 0.005679326975465212, "darwin": 0.9761200422513723, "sagan": 0.004748419008281577, "ibn_sina": 0.0036160960581707842, "einstein": 0.00608916201835754, "socrates": 0.0037469536883525823}, {"ibn_sina": 0.005797916861768259, "einstein": 0.014564223811048858, "darwin": 0.962004549970995, "sagan": 0.006305092476385611, "hypatia": 0.003946629189335062, "socrates": 0.007381587690467132}, {"einstein": 0.00604699140716501, "sagan": 0.0050782028123968, "socrates": 0.0037229203390477597, "ibn_sina": 0.0038962430413819705, "darwin": 0.9753303456787976, "hypatia": 0.005925296721210862}, {"socrates": 0.0038910919604199117, "ibn_sina": 0.0026708132079742222, "hypatia": 0.002483629134805106, "einstein": 0.004306309673495991, "sagan": 0.005143500663795292, "darwin": 0.9815046553595095}], "prompt_tokens": 445, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.977843288450736, "einstein": 0.006797375120892521, "hypatia": 0.0033080956404960573, "ibn_sina": 0.0028704981224578546, "sagan": 0.004687513485026962, "socrates": 0.004493229180390756}, "confidence": 0.947542055468873}}, "usage": {"input_tokens": 445, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.0030862638873911975, "hypatia": 0.002630196393873711, "sagan": 0.004743679755755736, "socrates": 0.004305065428196521, "darwin": 0.9727325689621946, "einstein": 0.012502225572588118}, "confidence": 0.9357069224894534}}, "usage": {"input_tokens": 445, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.005679326975465212, "darwin": 0.9761200422513723, "sagan": 0.004748419008281577, "ibn_sina": 0.0036160960581707842, "einstein": 0.00608916201835754, "socrates": 0.0037469536883525823}, "confidence": 0.9435151993954284}}, "usage": {"input_tokens": 445, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.005797916861768259, "einstein": 0.014564223811048858, "darwin": 0.962004549970995, "sagan": 0.006305092476385611, "hypatia": 0.003946629189335062, "socrates": 0.007381587690467132}, "confidence": 0.9109699654359723}}, "usage": {"input_tokens": 445, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.00604699140716501, "sagan": 0.0050782028123968, "socrates": 0.0037229203390477597, "ibn_sina": 0.0038962430413819705, "darwin": 0.9753303456787976, "hypatia": 0.005925296721210862}, "confidence": 0.9416749449252734}}, "usage": {"input_tokens": 445, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.0038910919604199117, "ibn_sina": 0.0026708132079742222, "hypatia": 0.002483629134805106, "einstein": 0.004306309673495991, "sagan": 0.005143500663795292, "darwin": 0.9815046553595095}, "confidence": 0.9561097967775851}}, "usage": {"input_tokens": 445, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:34Z"}
{"id": "c-darwin-01-starter-01", "uid": "c014", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "prompt_sha256_by_order": ["9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564"], "choice": "sagan", "probs": {"darwin": 0.029792215777398733, "einstein": 0.01493447613686718, "hypatia": 0.009565749548255686, "ibn_sina": 0.00955100322017092, "sagan": 0.9087072056889336, "socrates": 0.027449349628373858}, "confidence": 0.7933547108696224, "seconds": 0.19154380116651737, "seconds_all_orders": 1.1492628069991042, "seconds_by_order": [0.19151473300007638, 0.19152881300033187, 0.19128394499966817, 0.19174582099913096, 0.19176812099976814, 0.1914213740001287], "argmax_stable": true, "spread": {"darwin": 0.017964660905997604, "einstein": 0.014486525599806375, "hypatia": 0.008088686296349578, "ibn_sina": 0.0028661411237825268, "sagan": 0.050420021790515634, "socrates": 0.02969918987199497}, "spread_2dp": {"darwin": 0.019999999999999997, "einstein": 0.019999999999999997, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.050000000000000044, "socrates": 0.030000000000000002}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.029792215777398733, "einstein": 0.01493447613686718, "hypatia": 0.009565749548255686, "ibn_sina": 0.00955100322017092, "sagan": 0.9087072056889336, "socrates": 0.027449349628373858}, {"ibn_sina": 0.00723268932792537, "hypatia": 0.013772438686154135, "sagan": 0.911732467123946, "socrates": 0.02112395461154612, "darwin": 0.016717448513754885, "einstein": 0.029421001736673554}, {"hypatia": 0.009275980552914492, "darwin": 0.012616600043409922, "sagan": 0.911750971151478, "ibn_sina": 0.008897845366090595, "einstein": 0.020239215214764378, "socrates": 0.03721938767134274}, {"ibn_sina": 0.008103764372654757, "einstein": 0.020793292305610306, "darwin": 0.017843374189057255, "sagan": 0.9113214093582681, "hypatia": 0.01062468538073112, "socrates": 0.03131347439367843}, {"einstein": 0.02074707595984032, "sagan": 0.9285746040067592, "socrates": 0.020159371317560584, "ibn_sina": 0.007455097170236496, "darwin": 0.01182755487140113, "hypatia": 0.011236296674202352}, {"socrates": 0.049858561189555554, "ibn_sina": 0.0066848620963883935, "hypatia": 0.01736466684926407, "einstein": 0.0266676275665894, "sagan": 0.8781545822162435, "darwin": 0.021269700081958998}], "prompt_tokens": 495, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.029792215777398733, "einstein": 0.01493447613686718, "hypatia": 0.009565749548255686, "ibn_sina": 0.00955100322017092, "sagan": 0.9087072056889336, "socrates": 0.027449349628373858}, "confidence": 0.7933547108696224}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.00723268932792537, "hypatia": 0.013772438686154135, "sagan": 0.911732467123946, "socrates": 0.02112395461154612, "darwin": 0.016717448513754885, "einstein": 0.029421001736673554}, "confidence": 0.7997072480279506}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.009275980552914492, "darwin": 0.012616600043409922, "sagan": 0.911750971151478, "ibn_sina": 0.008897845366090595, "einstein": 0.020239215214764378, "socrates": 0.03721938767134274}, "confidence": 0.8000909633330018}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.008103764372654757, "einstein": 0.020793292305610306, "darwin": 0.017843374189057255, "sagan": 0.9113214093582681, "hypatia": 0.01062468538073112, "socrates": 0.03131347439367843}, "confidence": 0.7988998561325775}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.02074707595984032, "sagan": 0.9285746040067592, "socrates": 0.020159371317560584, "ibn_sina": 0.007455097170236496, "darwin": 0.01182755487140113, "hypatia": 0.011236296674202352}, "confidence": 0.8360912326121399}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.049858561189555554, "ibn_sina": 0.0066848620963883935, "hypatia": 0.01736466684926407, "einstein": 0.0266676275665894, "sagan": 0.8781545822162435, "darwin": 0.021269700081958998}, "confidence": 0.7301813535147021}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:35Z"}
{"id": "c-darwin-01-starter-02", "uid": "c015", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "prompt_sha256_by_order": ["30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb"], "choice": "darwin", "probs": {"darwin": 0.8957077409049548, "einstein": 0.04168760180784761, "hypatia": 0.005863724932069082, "ibn_sina": 0.004797658470478125, "sagan": 0.03450995773837077, "socrates": 0.017433316146279708}, "confidence": 0.766698966104063, "seconds": 0.18766969266653177, "seconds_all_orders": 1.1260181559991906, "seconds_by_order": [0.18777731299996958, 0.18731170699902577, 0.18762151500050095, 0.18795411200153467, 0.18744652599889378, 0.1879069829992659], "argmax_stable": true, "spread": {"darwin": 0.0840685962140173, "einstein": 0.07589249771135825, "hypatia": 0.004248558465882405, "ibn_sina": 0.0033513451181100273, "sagan": 0.019951307051786195, "socrates": 0.029928541539705}, "spread_2dp": {"darwin": 0.08000000000000007, "einstein": 0.07, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.019999999999999997, "socrates": 0.03}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.8957077409049548, "einstein": 0.04168760180784761, "hypatia": 0.005863724932069082, "ibn_sina": 0.004797658470478125, "sagan": 0.03450995773837077, "socrates": 0.017433316146279708}, {"ibn_sina": 0.0052003899079556895, "hypatia": 0.005471280900045127, "sagan": 0.019917718780290604, "socrates": 0.009325875921743117, "darwin": 0.9248070356249023, "einstein": 0.035277698865063115}, {"hypatia": 0.0071264325970085, "darwin": 0.840738439410885, "sagan": 0.019705914477058044, "ibn_sina": 0.00508123295680959, "einstein": 0.11117019657642137, "socrates": 0.01617778398181742}, {"ibn_sina": 0.008149003588588152, "einstein": 0.036254923062105024, "darwin": 0.8954591772347258, "sagan": 0.014558650686584573, "hypatia": 0.006323827966548267, "socrates": 0.039254417461448116}, {"einstein": 0.05706140912412472, "sagan": 0.02806314619738655, "socrates": 0.010523504929053224, "ibn_sina": 0.00663173821286091, "darwin": 0.8889511351160526, "hypatia": 0.00876906642052188}, {"socrates": 0.023092681510713704, "ibn_sina": 0.006562613693215216, "hypatia": 0.004520507954639475, "einstein": 0.05592710930589505, "sagan": 0.02743283398187201, "darwin": 0.8824642535536644}], "prompt_tokens": 453, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.8957077409049548, "einstein": 0.04168760180784761, "hypatia": 0.005863724932069082, "ibn_sina": 0.004797658470478125, "sagan": 0.03450995773837077, "socrates": 0.017433316146279708}, "confidence": 0.766698966104063}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.0052003899079556895, "hypatia": 0.005471280900045127, "sagan": 0.019917718780290604, "socrates": 0.009325875921743117, "darwin": 0.9248070356249023, "einstein": 0.035277698865063115}, "confidence": 0.8284638827578406}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.0071264325970085, "darwin": 0.840738439410885, "sagan": 0.019705914477058044, "ibn_sina": 0.00508123295680959, "einstein": 0.11117019657642137, "socrates": 0.01617778398181742}, "confidence": 0.6639119018077014}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.008149003588588152, "einstein": 0.036254923062105024, "darwin": 0.8954591772347258, "sagan": 0.014558650686584573, "hypatia": 0.006323827966548267, "socrates": 0.039254417461448116}, "confidence": 0.7660249818397842}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.05706140912412472, "sagan": 0.02806314619738655, "socrates": 0.010523504929053224, "ibn_sina": 0.00663173821286091, "darwin": 0.8889511351160526, "hypatia": 0.00876906642052188}, "confidence": 0.7534111430121382}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.023092681510713704, "ibn_sina": 0.006562613693215216, "hypatia": 0.004520507954639475, "einstein": 0.05592710930589505, "sagan": 0.02743283398187201, "darwin": 0.8824642535536644}, "confidence": 0.7398644026787515}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:36Z"}
{"id": "c-darwin-01-starter-03", "uid": "c016", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "prompt_sha256_by_order": ["e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8"], "choice": "sagan", "probs": {"darwin": 0.2885072729374351, "einstein": 0.03162472526517588, "hypatia": 0.031234225028287183, "ibn_sina": 0.03584987704409367, "sagan": 0.5879151586140627, "socrates": 0.024868741110945496}, "confidence": 0.3193120579548755, "seconds": 0.2529755801670035, "seconds_all_orders": 1.517853481002021, "seconds_by_order": [0.2530166670003382, 0.252814658000716, 0.2530997459998616, 0.25237930200091796, 0.25324556399937137, 0.25329754400081583], "argmax_stable": true, "spread": {"darwin": 0.23726170369602917, "einstein": 0.01212795379654237, "hypatia": 0.03609044727659494, "ibn_sina": 0.018781026789889615, "sagan": 0.23582591174713352, "socrates": 0.007685140610343304}, "spread_2dp": {"darwin": 0.24000000000000002, "einstein": 0.010000000000000002, "hypatia": 0.039999999999999994, "ibn_sina": 0.02, "sagan": 0.22999999999999998, "socrates": 0.009999999999999998}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.2885072729374351, "einstein": 0.03162472526517588, "hypatia": 0.031234225028287183, "ibn_sina": 0.03584987704409367, "sagan": 0.5879151586140627, "socrates": 0.024868741110945496}, {"ibn_sina": 0.017068850254204054, "hypatia": 0.03385100010648452, "sagan": 0.7916647660268443, "socrates": 0.02679048168923656, "darwin": 0.10087295302799099, "einstein": 0.029751948895239536}, {"hypatia": 0.023860593830832368, "darwin": 0.33813465672402016, "sagan": 0.5558388542797108, "ibn_sina": 0.032143729586399855, "einstein": 0.029235586952774853, "socrates": 0.02078657862626208}, {"ibn_sina": 0.02177948012017978, "einstein": 0.03290791846364347, "darwin": 0.1387968774350227, "sagan": 0.762461201306514, "hypatia": 0.02164269974421411, "socrates": 0.02241182293042599}, {"einstein": 0.02743468467804995, "sagan": 0.6777186182421306, "socrates": 0.02735262436909918, "ibn_sina": 0.03218995087291517, "darwin": 0.2159693903815302, "hypatia": 0.01933473145627484}, {"socrates": 0.019667483758755877, "ibn_sina": 0.02183966863777402, "hypatia": 0.055425178732869784, "einstein": 0.03956263847459232, "sagan": 0.747521212981445, "darwin": 0.11598381741456296}], "prompt_tokens": 697, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.2885072729374351, "einstein": 0.03162472526517588, "hypatia": 0.031234225028287183, "ibn_sina": 0.03584987704409367, "sagan": 0.5879151586140627, "socrates": 0.024868741110945496}, "confidence": 0.3193120579548755}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.017068850254204054, "hypatia": 0.03385100010648452, "sagan": 0.7916647660268443, "socrates": 0.02679048168923656, "darwin": 0.10087295302799099, "einstein": 0.029751948895239536}, "confidence": 0.5679383183804096}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.023860593830832368, "darwin": 0.33813465672402016, "sagan": 0.5558388542797108, "ibn_sina": 0.032143729586399855, "einstein": 0.029235586952774853, "socrates": 0.02078657862626208}, "confidence": 0.3114174720279942}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.02177948012017978, "einstein": 0.03290791846364347, "darwin": 0.1387968774350227, "sagan": 0.762461201306514, "hypatia": 0.02164269974421411, "socrates": 0.02241182293042599}, "confidence": 0.52376755575362}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.02743468467804995, "sagan": 0.6777186182421306, "socrates": 0.02735262436909918, "ibn_sina": 0.03218995087291517, "darwin": 0.2159693903815302, "hypatia": 0.01933473145627484}, "confidence": 0.4106273870256888}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.019667483758755877, "ibn_sina": 0.02183966863777402, "hypatia": 0.055425178732869784, "einstein": 0.03956263847459232, "sagan": 0.747521212981445, "darwin": 0.11598381741456296}, "confidence": 0.49328937232426745}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:38Z"}
{"id": "c-darwin-01-starter-03", "uid": "c017", "task": "kperm", "label": "darwin", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "prompt_sha256_by_order": ["d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169"], "choice": "darwin", "probs": {"darwin": 0.7776828062852583, "einstein": 0.1002684319427464, "hypatia": 0.012567029223741633, "ibn_sina": 0.06293493456872731, "sagan": 0.02104849647143539, "socrates": 0.02549830150809097}, "confidence": 0.5440674933185893, "seconds": 0.22869499533347457, "seconds_all_orders": 1.3721699720008473, "seconds_by_order": [0.22856125799989968, 0.22813823199976468, 0.2287076070006151, 0.2290952940002171, 0.22881460600001446, 0.2288529750003363], "argmax_stable": true, "spread": {"darwin": 0.09008002436518914, "einstein": 0.046011194949011146, "hypatia": 0.004816861551867649, "ibn_sina": 0.03306842421857998, "sagan": 0.022294650394530817, "socrates": 0.016589741648973196}, "spread_2dp": {"darwin": 0.08999999999999997, "einstein": 0.05, "hypatia": 0.0, "ibn_sina": 0.03, "sagan": 0.019999999999999997, "socrates": 0.019999999999999997}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.7776828062852583, "einstein": 0.1002684319427464, "hypatia": 0.012567029223741633, "ibn_sina": 0.06293493456872731, "sagan": 0.02104849647143539, "socrates": 0.02549830150809097}, {"ibn_sina": 0.029866510350147334, "hypatia": 0.010454419465106739, "sagan": 0.015329988872395835, "socrates": 0.013604325364614495, "darwin": 0.8677628306504475, "einstein": 0.06298192529728816}, {"hypatia": 0.01424552508236493, "darwin": 0.847365931451906, "sagan": 0.024511167307048002, "ibn_sina": 0.03497608691959391, "einstein": 0.060764997830834104, "socrates": 0.01813629140825306}, {"ibn_sina": 0.04386188796054255, "einstein": 0.07711253507065757, "darwin": 0.8241574115829224, "sagan": 0.012297699384363316, "hypatia": 0.012376398987926487, "socrates": 0.03019406701358769}, {"einstein": 0.0690264541769891, "sagan": 0.019693122088548863, "socrates": 0.013923660342694327, "ibn_sina": 0.030019252960305075, "darwin": 0.8530007908542327, "hypatia": 0.01433671957722998}, {"socrates": 0.01943753013986913, "ibn_sina": 0.05542050718531007, "hypatia": 0.009519858025362331, "einstein": 0.054257236993735254, "sagan": 0.03459234977889413, "darwin": 0.826772517876829}], "prompt_tokens": 669, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.7776828062852583, "einstein": 0.1002684319427464, "hypatia": 0.012567029223741633, "ibn_sina": 0.06293493456872731, "sagan": 0.02104849647143539, "socrates": 0.02549830150809097}, "confidence": 0.5440674933185893}}, "usage": {"input_tokens": 669, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.029866510350147334, "hypatia": 0.010454419465106739, "sagan": 0.015329988872395835, "socrates": 0.013604325364614495, "darwin": 0.8677628306504475, "einstein": 0.06298192529728816}, "confidence": 0.7100805312723187}}, "usage": {"input_tokens": 669, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.01424552508236493, "darwin": 0.847365931451906, "sagan": 0.024511167307048002, "ibn_sina": 0.03497608691959391, "einstein": 0.060764997830834104, "socrates": 0.01813629140825306}, "confidence": 0.6688928689319391}}, "usage": {"input_tokens": 669, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.04386188796054255, "einstein": 0.07711253507065757, "darwin": 0.8241574115829224, "sagan": 0.012297699384363316, "hypatia": 0.012376398987926487, "socrates": 0.03019406701358769}, "confidence": 0.625986085230912}}, "usage": {"input_tokens": 669, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.0690264541769891, "sagan": 0.019693122088548863, "socrates": 0.013923660342694327, "ibn_sina": 0.030019252960305075, "darwin": 0.8530007908542327, "hypatia": 0.01433671957722998}, "confidence": 0.6808760620307904}}, "usage": {"input_tokens": 669, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.01943753013986913, "ibn_sina": 0.05542050718531007, "hypatia": 0.009519858025362331, "einstein": 0.054257236993735254, "sagan": 0.03459234977889413, "darwin": 0.826772517876829}, "confidence": 0.6294797831646995}}, "usage": {"input_tokens": 669, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:39Z"}
{"id": "c-einstein-01-dessert-01", "uid": "c018", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "prompt_sha256_by_order": ["f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3"], "choice": "sagan", "probs": {"darwin": 0.031991386576434556, "einstein": 0.06894013849201672, "hypatia": 0.05459557831662985, "ibn_sina": 0.03247359337982634, "sagan": 0.7450142936868319, "socrates": 0.06698500954826067}, "confidence": 0.48321363070117174, "seconds": 0.22589085316697796, "seconds_all_orders": 1.3553451190018677, "seconds_by_order": [0.22579629100073362, 0.22587034999924072, 0.22644436500013398, 0.22576923100132262, 0.22586025000055088, 0.2256046319998859], "argmax_stable": true, "spread": {"darwin": 0.02202141611959122, "einstein": 0.05043578717044943, "hypatia": 0.024776248065942225, "ibn_sina": 0.013905046974122086, "sagan": 0.05827337513266484, "socrates": 0.023035508138907063}, "spread_2dp": {"darwin": 0.02, "einstein": 0.04999999999999999, "hypatia": 0.03, "ibn_sina": 0.02, "sagan": 0.06000000000000005, "socrates": 0.01999999999999999}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.031991386576434556, "einstein": 0.06894013849201672, "hypatia": 0.05459557831662985, "ibn_sina": 0.03247359337982634, "sagan": 0.7450142936868319, "socrates": 0.06698500954826067}, {"ibn_sina": 0.025576028507440243, "hypatia": 0.05211912983298962, "sagan": 0.709732080582421, "socrates": 0.06811649713131157, "darwin": 0.02508033828337139, "einstein": 0.11937592566246615}, {"hypatia": 0.06386755622287303, "darwin": 0.029439388883809967, "sagan": 0.6901769349797043, "ibn_sina": 0.03038868334354681, "einstein": 0.09662901928135434, "socrates": 0.08949841728871158}, {"ibn_sina": 0.03586456828826618, "einstein": 0.09439531198746401, "darwin": 0.035085164067780804, "sagan": 0.6867409185541671, "hypatia": 0.05789351941515414, "socrates": 0.09002051768716773}, {"einstein": 0.07735228342951758, "sagan": 0.7380179407838368, "socrates": 0.06776757278887456, "ibn_sina": 0.021959521314144095, "darwin": 0.021048407441078713, "hypatia": 0.07385427424254833}, {"socrates": 0.07272622848194223, "ibn_sina": 0.02601849666162438, "hypatia": 0.07689537789893185, "einstein": 0.07783032948988032, "sagan": 0.7034597439069512, "darwin": 0.043069823560669934}], "prompt_tokens": 649, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.031991386576434556, "einstein": 0.06894013849201672, "hypatia": 0.05459557831662985, "ibn_sina": 0.03247359337982634, "sagan": 0.7450142936868319, "socrates": 0.06698500954826067}, "confidence": 0.48321363070117174}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.025576028507440243, "hypatia": 0.05211912983298962, "sagan": 0.709732080582421, "socrates": 0.06811649713131157, "darwin": 0.02508033828337139, "einstein": 0.11937592566246615}, "confidence": 0.43193158637707907}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.06386755622287303, "darwin": 0.029439388883809967, "sagan": 0.6901769349797043, "ibn_sina": 0.03038868334354681, "einstein": 0.09662901928135434, "socrates": 0.08949841728871158}, "confidence": 0.3994726600882338}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.03586456828826618, "einstein": 0.09439531198746401, "darwin": 0.035085164067780804, "sagan": 0.6867409185541671, "hypatia": 0.05789351941515414, "socrates": 0.09002051768716773}, "confidence": 0.39339538399952456}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.07735228342951758, "sagan": 0.7380179407838368, "socrates": 0.06776757278887456, "ibn_sina": 0.021959521314144095, "darwin": 0.021048407441078713, "hypatia": 0.07385427424254833}, "confidence": 0.4739512125379794}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.07272622848194223, "ibn_sina": 0.02601849666162438, "hypatia": 0.07689537789893185, "einstein": 0.07783032948988032, "sagan": 0.7034597439069512, "darwin": 0.043069823560669934}, "confidence": 0.41757657616917054}}, "usage": {"input_tokens": 649, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:41Z"}
{"id": "c-einstein-01-dessert-01", "uid": "c019", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "prompt_sha256_by_order": ["b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd"], "choice": "sagan", "probs": {"darwin": 0.005220794612717882, "einstein": 0.04783958040792127, "hypatia": 0.004083669336765081, "ibn_sina": 0.00416152051727037, "sagan": 0.9342011354041735, "socrates": 0.0044932997211517955}, "confidence": 0.8501221934693113, "seconds": 0.22772167633350668, "seconds_all_orders": 1.36633005800104, "seconds_by_order": [0.22804184200140298, 0.2279317730008188, 0.22743309699944803, 0.2269288109982881, 0.22797745300158567, 0.22801708199949644], "argmax_stable": true, "spread": {"darwin": 0.003939558452236435, "einstein": 0.06263013420496368, "hypatia": 0.002217068508452904, "ibn_sina": 0.002833530211502614, "sagan": 0.06784378260797741, "socrates": 0.002607895556707536}, "spread_2dp": {"darwin": 0.01, "einstein": 0.06999999999999999, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.06999999999999995, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.005220794612717882, "einstein": 0.04783958040792127, "hypatia": 0.004083669336765081, "ibn_sina": 0.00416152051727037, "sagan": 0.9342011354041735, "socrates": 0.0044932997211517955}, {"ibn_sina": 0.0036218388549414506, "hypatia": 0.004713633574985804, "sagan": 0.9288510467101965, "socrates": 0.005438599504103595, "darwin": 0.005580002353763939, "einstein": 0.05179487900200876}, {"hypatia": 0.004350945869540178, "darwin": 0.005213341625798628, "sagan": 0.8917301951931387, "ibn_sina": 0.005759103331595784, "einstein": 0.08742401705376171, "socrates": 0.005522396926165009}, {"ibn_sina": 0.0052423140791657335, "einstein": 0.07255337163542323, "darwin": 0.006818790888468036, "sagan": 0.9059101483688538, "hypatia": 0.004997492185036817, "socrates": 0.004477882843052486}, {"einstein": 0.03158707008330368, "sagan": 0.9553615700332347, "socrates": 0.0044661306505528395, "ibn_sina": 0.0029255731200931697, "darwin": 0.0028792324362316006, "hypatia": 0.0027804236765839128}, {"socrates": 0.002914501369457473, "ibn_sina": 0.002985292551136339, "hypatia": 0.003870554033904794, "einstein": 0.02479388284879804, "sagan": 0.9595739778011161, "darwin": 0.005861791395587223}], "prompt_tokens": 663, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.005220794612717882, "einstein": 0.04783958040792127, "hypatia": 0.004083669336765081, "ibn_sina": 0.00416152051727037, "sagan": 0.9342011354041735, "socrates": 0.0044932997211517955}, "confidence": 0.8501221934693113}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.0036218388549414506, "hypatia": 0.004713633574985804, "sagan": 0.9288510467101965, "socrates": 0.005438599504103595, "darwin": 0.005580002353763939, "einstein": 0.05179487900200876}, "confidence": 0.8386516327773578}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.004350945869540178, "darwin": 0.005213341625798628, "sagan": 0.8917301951931387, "ibn_sina": 0.005759103331595784, "einstein": 0.08742401705376171, "socrates": 0.005522396926165009}, "confidence": 0.7635225682922447}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.0052423140791657335, "einstein": 0.07255337163542323, "darwin": 0.006818790888468036, "sagan": 0.9059101483688538, "hypatia": 0.004997492185036817, "socrates": 0.004477882843052486}, "confidence": 0.7912674313387928}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.03158707008330368, "sagan": 0.9553615700332347, "socrates": 0.0044661306505528395, "ibn_sina": 0.0029255731200931697, "darwin": 0.0028792324362316006, "hypatia": 0.0027804236765839128}, "confidence": 0.896509598234952}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.002914501369457473, "ibn_sina": 0.002985292551136339, "hypatia": 0.003870554033904794, "einstein": 0.02479388284879804, "sagan": 0.9595739778011161, "darwin": 0.005861791395587223}, "confidence": 0.9057564442918238}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:42Z"}
{"id": "c-einstein-01-dessert-02", "uid": "c020", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "prompt_sha256_by_order": ["ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf"], "choice": "einstein", "probs": {"darwin": 0.03043946075429465, "einstein": 0.5984925298067715, "hypatia": 0.07592595037779379, "ibn_sina": 0.023488631887681383, "sagan": 0.05880930245232763, "socrates": 0.2128441247211311}, "confidence": 0.29703698830897773, "seconds": 0.221867754666467, "seconds_all_orders": 1.331206527998802, "seconds_by_order": [0.22180533400023705, 0.22184693300005165, 0.22148506599842221, 0.22173378399929788, 0.22184812300110934, 0.22248728799968376], "argmax_stable": true, "spread": {"darwin": 0.030090003979087596, "einstein": 0.15251747523201797, "hypatia": 0.25317467298016494, "ibn_sina": 0.10105098432855654, "sagan": 0.08808524290537592, "socrates": 0.2275453684358623}, "spread_2dp": {"darwin": 0.030000000000000002, "einstein": 0.14999999999999997, "hypatia": 0.25, "ibn_sina": 0.09999999999999999, "sagan": 0.09000000000000001, "socrates": 0.23}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.03043946075429465, "einstein": 0.5984925298067715, "hypatia": 0.07592595037779379, "ibn_sina": 0.023488631887681383, "sagan": 0.05880930245232763, "socrates": 0.2128441247211311}, {"ibn_sina": 0.11711342053225585, "hypatia": 0.2539814808260436, "sagan": 0.04854134937264012, "socrates": 0.028775728388549637, "darwin": 0.024479080934513606, "einstein": 0.5271089399459972}, {"hypatia": 0.29863596641585716, "darwin": 0.027461921916949357, "sagan": 0.07543668716521028, "ibn_sina": 0.016062436203699307, "einstein": 0.5093889199375423, "socrates": 0.07301406836074154}, {"ibn_sina": 0.09260511118947204, "einstein": 0.49346662031922806, "darwin": 0.01699264921675649, "sagan": 0.05162733806936563, "hypatia": 0.1249058191823329, "socrates": 0.2204024620228449}, {"einstein": 0.5389866858379652, "sagan": 0.05648066103539265, "socrates": 0.02804573169682837, "ibn_sina": 0.019257479301809995, "darwin": 0.028128818770045153, "hypatia": 0.3291006233579587}, {"socrates": 0.25559110013269065, "ibn_sina": 0.02339645291593092, "hypatia": 0.09132814690276486, "einstein": 0.4459750545747535, "sagan": 0.13662659227801605, "darwin": 0.047082653195844086}], "prompt_tokens": 605, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.03043946075429465, "einstein": 0.5984925298067715, "hypatia": 0.07592595037779379, "ibn_sina": 0.023488631887681383, "sagan": 0.05880930245232763, "socrates": 0.2128441247211311}, "confidence": 0.29703698830897773}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.11711342053225585, "hypatia": 0.2539814808260436, "sagan": 0.04854134937264012, "socrates": 0.028775728388549637, "darwin": 0.024479080934513606, "einstein": 0.5271089399459972}, "confidence": 0.23181941318674548}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.29863596641585716, "darwin": 0.027461921916949357, "sagan": 0.07543668716521028, "ibn_sina": 0.016062436203699307, "einstein": 0.5093889199375423, "socrates": 0.07301406836074154}, "confidence": 0.23283330298418586}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.09260511118947204, "einstein": 0.49346662031922806, "darwin": 0.01699264921675649, "sagan": 0.05162733806936563, "hypatia": 0.1249058191823329, "socrates": 0.2204024620228449}, "confidence": 0.18306142369908301}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.5389866858379652, "sagan": 0.05648066103539265, "socrates": 0.02804573169682837, "ibn_sina": 0.019257479301809995, "darwin": 0.028128818770045153, "hypatia": 0.3291006233579587}, "confidence": 0.2847430922764641}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.25559110013269065, "ibn_sina": 0.02339645291593092, "hypatia": 0.09132814690276486, "einstein": 0.4459750545747535, "sagan": 0.13662659227801605, "darwin": 0.047082653195844086}, "confidence": 0.1527908633742366}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:43Z"}
{"id": "c-einstein-01-dessert-02", "uid": "c021", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "prompt_sha256_by_order": ["dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516"], "choice": "einstein", "probs": {"darwin": 0.17888764294747203, "einstein": 0.5359264147901315, "hypatia": 0.018987674945884565, "ibn_sina": 0.01981305690150604, "sagan": 0.09346292538829956, "socrates": 0.15292228502670627}, "confidence": 0.2225098522872006, "seconds": 0.1905685159999848, "seconds_all_orders": 1.1434110959999089, "seconds_by_order": [0.19066283000029216, 0.18995953600096982, 0.19092024799829233, 0.190769339000326, 0.19095747800020035, 0.1901416649998282], "argmax_stable": true, "spread": {"darwin": 0.2107456357316665, "einstein": 0.19736583968980648, "hypatia": 0.02132933103944567, "ibn_sina": 0.1292757151873571, "sagan": 0.0653979165675189, "socrates": 0.25884580536760177}, "spread_2dp": {"darwin": 0.21000000000000002, "einstein": 0.2, "hypatia": 0.02, "ibn_sina": 0.13, "sagan": 0.06, "socrates": 0.25}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.17888764294747203, "einstein": 0.5359264147901315, "hypatia": 0.018987674945884565, "ibn_sina": 0.01981305690150604, "sagan": 0.09346292538829956, "socrates": 0.15292228502670627}, {"ibn_sina": 0.14908877208886312, "hypatia": 0.039172721401530886, "sagan": 0.02806500882078066, "socrates": 0.08511623907723065, "darwin": 0.2040513968328213, "einstein": 0.4945058617787734}, {"hypatia": 0.024085470956769325, "darwin": 0.2607664173101308, "sagan": 0.08368436654575616, "ibn_sina": 0.030175232898255175, "einstein": 0.5060267807255231, "socrates": 0.09526173156356549}, {"ibn_sina": 0.023993654713824086, "einstein": 0.48829898072367817, "darwin": 0.07130543769161522, "sagan": 0.04395325914750538, "hypatia": 0.02848662327854473, "socrates": 0.3439620444448324}, {"einstein": 0.5329844916121633, "sagan": 0.0801647131497272, "socrates": 0.10463725996353158, "ibn_sina": 0.03374596329760641, "darwin": 0.21623617898231415, "hypatia": 0.032231392994657285}, {"socrates": 0.23081293836568456, "ibn_sina": 0.045211108820573505, "hypatia": 0.017843390362085217, "einstein": 0.338560575100325, "sagan": 0.08552091392805002, "darwin": 0.28205107342328173}], "prompt_tokens": 481, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.17888764294747203, "einstein": 0.5359264147901315, "hypatia": 0.018987674945884565, "ibn_sina": 0.01981305690150604, "sagan": 0.09346292538829956, "socrates": 0.15292228502670627}, "confidence": 0.2225098522872006}}, "usage": {"input_tokens": 481, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.14908877208886312, "hypatia": 0.039172721401530886, "sagan": 0.02806500882078066, "socrates": 0.08511623907723065, "darwin": 0.2040513968328213, "einstein": 0.4945058617787734}, "confidence": 0.18156088338726714}}, "usage": {"input_tokens": 481, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.024085470956769325, "darwin": 0.2607664173101308, "sagan": 0.08368436654575616, "ibn_sina": 0.030175232898255175, "einstein": 0.5060267807255231, "socrates": 0.09526173156356549}, "confidence": 0.2099569030054294}}, "usage": {"input_tokens": 481, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.023993654713824086, "einstein": 0.48829898072367817, "darwin": 0.07130543769161522, "sagan": 0.04395325914750538, "hypatia": 0.02848662327854473, "socrates": 0.3439620444448324}, "confidence": 0.23817918424104692}}, "usage": {"input_tokens": 481, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.5329844916121633, "sagan": 0.0801647131497272, "socrates": 0.10463725996353158, "ibn_sina": 0.03374596329760641, "darwin": 0.21623617898231415, "hypatia": 0.032231392994657285}, "confidence": 0.22046025224831292}}, "usage": {"input_tokens": 481, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.23081293836568456, "ibn_sina": 0.045211108820573505, "hypatia": 0.017843390362085217, "einstein": 0.338560575100325, "sagan": 0.08552091392805002, "darwin": 0.28205107342328173}, "confidence": 0.10855232944957365}}, "usage": {"input_tokens": 481, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:44Z"}
{"id": "c-einstein-01-dessert-03", "uid": "c022", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "prompt_sha256_by_order": ["78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c"], "choice": "einstein", "probs": {"darwin": 0.12972381372136915, "einstein": 0.8410552848819011, "hypatia": 0.004614041134908429, "ibn_sina": 0.008370489217101354, "sagan": 0.006541209749426128, "socrates": 0.00969516129529382}, "confidence": 0.6693164777404579, "seconds": 0.25459365666696004, "seconds_all_orders": 1.5275619400017604, "seconds_by_order": [0.2542695160009316, 0.2546445930001937, 0.25474579200090375, 0.2542983860003005, 0.2546696930003236, 0.2549339599991072], "argmax_stable": true, "spread": {"darwin": 0.3300090902561252, "einstein": 0.3210034701308129, "hypatia": 0.0022957858144875345, "ibn_sina": 0.010730301982114145, "sagan": 0.005094523598603824, "socrates": 0.009747851855713658}, "spread_2dp": {"darwin": 0.33, "einstein": 0.31999999999999995, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.12972381372136915, "einstein": 0.8410552848819011, "hypatia": 0.004614041134908429, "ibn_sina": 0.008370489217101354, "sagan": 0.006541209749426128, "socrates": 0.00969516129529382}, {"ibn_sina": 0.008211713638862389, "hypatia": 0.004055156493386225, "sagan": 0.003342990130329993, "socrates": 0.0046054210088389775, "darwin": 0.45973290397749433, "einstein": 0.5200518147510882}, {"hypatia": 0.004656833874177565, "darwin": 0.4023090387270174, "sagan": 0.004466692429520642, "ibn_sina": 0.007008953396906318, "einstein": 0.5724401571165193, "socrates": 0.00911832445585878}, {"ibn_sina": 0.015605588923652308, "einstein": 0.563979693815753, "darwin": 0.3947256987297644, "sagan": 0.004984803358403999, "hypatia": 0.0063509423078737595, "socrates": 0.014353272864552635}, {"einstein": 0.5778581895830998, "sagan": 0.008437513728933817, "socrates": 0.007555038013513599, "ibn_sina": 0.0072309482493090015, "darwin": 0.3938203500308832, "hypatia": 0.00509796039426058}, {"socrates": 0.01202384368951329, "ibn_sina": 0.017739255379020463, "hypatia": 0.004285402816008072, "einstein": 0.6291590577112449, "sagan": 0.007789447960641719, "darwin": 0.3290029924435715}], "prompt_tokens": 715, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.12972381372136915, "einstein": 0.8410552848819011, "hypatia": 0.004614041134908429, "ibn_sina": 0.008370489217101354, "sagan": 0.006541209749426128, "socrates": 0.00969516129529382}, "confidence": 0.6693164777404579}}, "usage": {"input_tokens": 715, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.008211713638862389, "hypatia": 0.004055156493386225, "sagan": 0.003342990130329993, "socrates": 0.0046054210088389775, "darwin": 0.45973290397749433, "einstein": 0.5200518147510882}, "confidence": 0.3783093940554794}}, "usage": {"input_tokens": 715, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.004656833874177565, "darwin": 0.4023090387270174, "sagan": 0.004466692429520642, "ibn_sina": 0.007008953396906318, "einstein": 0.5724401571165193, "socrates": 0.00911832445585878}, "confidence": 0.3876570433991632}}, "usage": {"input_tokens": 715, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.015605588923652308, "einstein": 0.563979693815753, "darwin": 0.3947256987297644, "sagan": 0.004984803358403999, "hypatia": 0.0063509423078737595, "socrates": 0.014353272864552635}, "confidence": 0.36927544702541126}}, "usage": {"input_tokens": 715, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.5778581895830998, "sagan": 0.008437513728933817, "socrates": 0.007555038013513599, "ibn_sina": 0.0072309482493090015, "darwin": 0.3938203500308832, "hypatia": 0.00509796039426058}, "confidence": 0.38706532170013874}}, "usage": {"input_tokens": 715, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.01202384368951329, "ibn_sina": 0.017739255379020463, "hypatia": 0.004285402816008072, "einstein": 0.6291590577112449, "sagan": 0.007789447960641719, "darwin": 0.3290029924435715}, "confidence": 0.4055468597346703}}, "usage": {"input_tokens": 715, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:46Z"}
{"id": "c-einstein-01-main-01", "uid": "c023", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "prompt_sha256_by_order": ["18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c"], "choice": "einstein", "probs": {"darwin": 0.14443369228820008, "einstein": 0.5153369801757488, "hypatia": 0.03951746690677821, "ibn_sina": 0.10101560829240533, "sagan": 0.07229478283964103, "socrates": 0.1274014694972265}, "confidence": 0.183588097563951, "seconds": 0.18498308366694496, "seconds_all_orders": 1.1098985020016698, "seconds_by_order": [0.18503172700002324, 0.18456200100081332, 0.18500525700073922, 0.18532365399914852, 0.18520422400069947, 0.18477163900024607], "argmax_stable": false, "spread": {"darwin": 0.2077541830019733, "einstein": 0.2542040151722027, "hypatia": 0.04616149424358164, "ibn_sina": 0.08814607486183675, "sagan": 0.07185524493238662, "socrates": 0.13543150073823362}, "spread_2dp": {"darwin": 0.21, "einstein": 0.26, "hypatia": 0.05, "ibn_sina": 0.09, "sagan": 0.07, "socrates": 0.13999999999999999}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.14443369228820008, "einstein": 0.5153369801757488, "hypatia": 0.03951746690677821, "ibn_sina": 0.10101560829240533, "sagan": 0.07229478283964103, "socrates": 0.1274014694972265}, {"ibn_sina": 0.151450884977486, "hypatia": 0.04505216433081973, "sagan": 0.14415002777202765, "socrates": 0.07476534020756301, "darwin": 0.11705335229650585, "einstein": 0.46752823041559777}, {"hypatia": 0.07643921224049086, "darwin": 0.1840932823112392, "sagan": 0.1356281625706053, "ibn_sina": 0.08140293764797704, "einstein": 0.4234578955195579, "socrates": 0.09897850971012961}, {"ibn_sina": 0.07200919197363138, "einstein": 0.46259858534155046, "darwin": 0.08792097923174493, "sagan": 0.1229243638042469, "hypatia": 0.04639504483479118, "socrates": 0.2081518348140351}, {"einstein": 0.4569268830204485, "sagan": 0.10481045323170671, "socrates": 0.07272033407580147, "ibn_sina": 0.08875792148535387, "darwin": 0.196993066963978, "hypatia": 0.07979134122271143}, {"socrates": 0.1422449981605265, "ibn_sina": 0.16015526683546813, "hypatia": 0.033629846979129796, "einstein": 0.2611329650035461, "sagan": 0.10716176078761144, "darwin": 0.2956751622337182}], "prompt_tokens": 505, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.14443369228820008, "einstein": 0.5153369801757488, "hypatia": 0.03951746690677821, "ibn_sina": 0.10101560829240533, "sagan": 0.07229478283964103, "socrates": 0.1274014694972265}, "confidence": 0.183588097563951}}, "usage": {"input_tokens": 505, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.151450884977486, "hypatia": 0.04505216433081973, "sagan": 0.14415002777202765, "socrates": 0.07476534020756301, "darwin": 0.11705335229650585, "einstein": 0.46752823041559777}, "confidence": 0.14034434583252886}}, "usage": {"input_tokens": 505, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.07643921224049086, "darwin": 0.1840932823112392, "sagan": 0.1356281625706053, "ibn_sina": 0.08140293764797704, "einstein": 0.4234578955195579, "socrates": 0.09897850971012961}, "confidence": 0.1046416733948349}}, "usage": {"input_tokens": 505, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.07200919197363138, "einstein": 0.46259858534155046, "darwin": 0.08792097923174493, "sagan": 0.1229243638042469, "hypatia": 0.04639504483479118, "socrates": 0.2081518348140351}, "confidence": 0.14500355105934215}}, "usage": {"input_tokens": 505, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.4569268830204485, "sagan": 0.10481045323170671, "socrates": 0.07272033407580147, "ibn_sina": 0.08875792148535387, "darwin": 0.196993066963978, "hypatia": 0.07979134122271143}, "confidence": 0.13372785965668993}}, "usage": {"input_tokens": 505, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.1422449981605265, "ibn_sina": 0.16015526683546813, "hypatia": 0.033629846979129796, "einstein": 0.2611329650035461, "sagan": 0.10716176078761144, "darwin": 0.2956751622337182}, "confidence": 0.056934622663897504}}, "usage": {"input_tokens": 505, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:47Z"}
{"id": "c-einstein-01-main-01", "uid": "c024", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "prompt_sha256_by_order": ["0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2"], "choice": "einstein", "probs": {"darwin": 0.013118118034673693, "einstein": 0.9452761936379177, "hypatia": 0.008208768970235331, "ibn_sina": 0.0078094055023563294, "sagan": 0.014517241002271978, "socrates": 0.011070272852545035}, "confidence": 0.87301700705082, "seconds": 0.2211999596665919, "seconds_all_orders": 1.3271997579995514, "seconds_by_order": [0.22090288999970653, 0.22112687899971206, 0.22120571800041944, 0.22181280299992068, 0.22130113700040965, 0.220850330999383], "argmax_stable": true, "spread": {"darwin": 0.008422189402947027, "einstein": 0.02312219823749395, "hypatia": 0.006795793172730508, "ibn_sina": 0.00817245844684963, "sagan": 0.013964800907185982, "socrates": 0.009605146968361273}, "spread_2dp": {"darwin": 0.01, "einstein": 0.029999999999999916, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.013118118034673693, "einstein": 0.9452761936379177, "hypatia": 0.008208768970235331, "ibn_sina": 0.0078094055023563294, "sagan": 0.014517241002271978, "socrates": 0.011070272852545035}, {"ibn_sina": 0.006878990917131846, "hypatia": 0.011921368055792286, "sagan": 0.008482409310918922, "socrates": 0.00962810427048436, "darwin": 0.015683279974561058, "einstein": 0.9474058474711116}, {"hypatia": 0.0063853732410933915, "darwin": 0.018598773778758432, "sagan": 0.009401245136805443, "ibn_sina": 0.009709017103284784, "einstein": 0.9398000430610717, "socrates": 0.016105547678986268}, {"ibn_sina": 0.010968903843465376, "einstein": 0.9450851678337457, "darwin": 0.010176584375811405, "sagan": 0.008728832139080562, "hypatia": 0.0131811664138239, "socrates": 0.011859345394073}, {"einstein": 0.9408840504373335, "sagan": 0.0165031056927898, "socrates": 0.010691531269961157, "ibn_sina": 0.006614103620610309, "darwin": 0.017885884029428922, "hypatia": 0.007421324949876266}, {"socrates": 0.019233251238845634, "ibn_sina": 0.01478656206745994, "hypatia": 0.008973109165871261, "einstein": 0.9242836492336176, "sagan": 0.022447210218104904, "darwin": 0.010276218076100688}], "prompt_tokens": 597, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.013118118034673693, "einstein": 0.9452761936379177, "hypatia": 0.008208768970235331, "ibn_sina": 0.0078094055023563294, "sagan": 0.014517241002271978, "socrates": 0.011070272852545035}, "confidence": 0.87301700705082}}, "usage": {"input_tokens": 597, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.006878990917131846, "hypatia": 0.011921368055792286, "sagan": 0.008482409310918922, "socrates": 0.00962810427048436, "darwin": 0.015683279974561058, "einstein": 0.9474058474711116}, "confidence": 0.8778134755421638}}, "usage": {"input_tokens": 597, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.0063853732410933915, "darwin": 0.018598773778758432, "sagan": 0.009401245136805443, "ibn_sina": 0.009709017103284784, "einstein": 0.9398000430610717, "socrates": 0.016105547678986268}, "confidence": 0.8608634144852416}}, "usage": {"input_tokens": 597, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.010968903843465376, "einstein": 0.9450851678337457, "darwin": 0.010176584375811405, "sagan": 0.008728832139080562, "hypatia": 0.0131811664138239, "socrates": 0.011859345394073}, "confidence": 0.8725605206945699}}, "usage": {"input_tokens": 597, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.9408840504373335, "sagan": 0.0165031056927898, "socrates": 0.010691531269961157, "ibn_sina": 0.006614103620610309, "darwin": 0.017885884029428922, "hypatia": 0.007421324949876266}, "confidence": 0.8632818219807958}}, "usage": {"input_tokens": 597, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.019233251238845634, "ibn_sina": 0.01478656206745994, "hypatia": 0.008973109165871261, "einstein": 0.9242836492336176, "sagan": 0.022447210218104904, "darwin": 0.010276218076100688}, "confidence": 0.8266945830450766}}, "usage": {"input_tokens": 597, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:48Z"}
{"id": "c-einstein-01-main-02", "uid": "c025", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "prompt_sha256_by_order": ["af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11"], "choice": "sagan", "probs": {"darwin": 0.016836209077069442, "einstein": 0.16605561336256386, "hypatia": 0.037232330702992776, "ibn_sina": 0.026062376463900035, "sagan": 0.6820459236955325, "socrates": 0.07176754669794147}, "confidence": 0.4003127696446727, "seconds": 0.22900306766678113, "seconds_all_orders": 1.3740184060006868, "seconds_by_order": [0.22883377600010135, 0.22929839199969138, 0.22917128300105105, 0.22880580600030953, 0.2285033680000197, 0.2294057809995138], "argmax_stable": true, "spread": {"darwin": 0.016081483689811514, "einstein": 0.16238526649624035, "hypatia": 0.01940427394210336, "ibn_sina": 0.008562492103135984, "sagan": 0.20757140464861978, "socrates": 0.05037515861238637}, "spread_2dp": {"darwin": 0.019999999999999997, "einstein": 0.16, "hypatia": 0.020000000000000004, "ibn_sina": 0.009999999999999998, "sagan": 0.20999999999999996, "socrates": 0.05}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.016836209077069442, "einstein": 0.16605561336256386, "hypatia": 0.037232330702992776, "ibn_sina": 0.026062376463900035, "sagan": 0.6820459236955325, "socrates": 0.07176754669794147}, {"ibn_sina": 0.025581280396653455, "hypatia": 0.047949600592411795, "sagan": 0.5009508820567049, "socrates": 0.08774612333324155, "darwin": 0.02928344879752326, "einstein": 0.30848866482346493}, {"hypatia": 0.040095953134403825, "darwin": 0.01884251847294675, "sagan": 0.618987624591353, "ibn_sina": 0.0224009608204963, "einstein": 0.18575046302698087, "socrates": 0.11392247995381931}, {"ibn_sina": 0.0249458920140753, "einstein": 0.26721249197624086, "darwin": 0.026904920723263226, "sagan": 0.5571608916800471, "hypatia": 0.03932718572570408, "socrates": 0.08444861788066937}, {"einstein": 0.14610339832722458, "sagan": 0.7085222867053247, "socrates": 0.06354732134143294, "ibn_sina": 0.01749988436076405, "darwin": 0.013201965107711745, "hypatia": 0.05112514415754212}, {"socrates": 0.06422677284319261, "ibn_sina": 0.020485314108250052, "hypatia": 0.03172087021543876, "einstein": 0.1655189540492905, "sagan": 0.6953861296122814, "darwin": 0.022661959171546564}], "prompt_tokens": 671, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.016836209077069442, "einstein": 0.16605561336256386, "hypatia": 0.037232330702992776, "ibn_sina": 0.026062376463900035, "sagan": 0.6820459236955325, "socrates": 0.07176754669794147}, "confidence": 0.4003127696446727}}, "usage": {"input_tokens": 671, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.025581280396653455, "hypatia": 0.047949600592411795, "sagan": 0.5009508820567049, "socrates": 0.08774612333324155, "darwin": 0.02928344879752326, "einstein": 0.30848866482346493}, "confidence": 0.22915301343417802}}, "usage": {"input_tokens": 671, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.040095953134403825, "darwin": 0.01884251847294675, "sagan": 0.618987624591353, "ibn_sina": 0.0224009608204963, "einstein": 0.18575046302698087, "socrates": 0.11392247995381931}, "confidence": 0.3197101292279836}}, "usage": {"input_tokens": 671, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.0249458920140753, "einstein": 0.26721249197624086, "darwin": 0.026904920723263226, "sagan": 0.5571608916800471, "hypatia": 0.03932718572570408, "socrates": 0.08444861788066937}, "confidence": 0.27022617276690414}}, "usage": {"input_tokens": 671, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.14610339832722458, "sagan": 0.7085222867053247, "socrates": 0.06354732134143294, "ibn_sina": 0.01749988436076405, "darwin": 0.013201965107711745, "hypatia": 0.05112514415754212}, "confidence": 0.43657913681326255}}, "usage": {"input_tokens": 671, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.06422677284319261, "ibn_sina": 0.020485314108250052, "hypatia": 0.03172087021543876, "einstein": 0.1655189540492905, "sagan": 0.6953861296122814, "darwin": 0.022661959171546564}, "confidence": 0.42042747742168257}}, "usage": {"input_tokens": 671, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:50Z"}
{"id": "c-einstein-01-main-03", "uid": "c026", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "prompt_sha256_by_order": ["fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd"], "choice": "sagan", "probs": {"darwin": 0.008531779010949772, "einstein": 0.011844572850719527, "hypatia": 0.006411547442246085, "ibn_sina": 0.0055551261976692105, "sagan": 0.9523236109174273, "socrates": 0.015333363580988076}, "confidence": 0.8889285093715579, "seconds": 0.22021345500009679, "seconds_all_orders": 1.3212807300005807, "seconds_by_order": [0.22052502400038065, 0.2202788859995053, 0.21996086900071532, 0.21998614800031646, 0.22027331599929312, 0.22025648700036982], "argmax_stable": true, "spread": {"darwin": 0.0037258213981906157, "einstein": 0.021399149446006835, "hypatia": 0.0035477600697476647, "ibn_sina": 0.0032680471665057394, "sagan": 0.0334215622114592, "socrates": 0.019783646129782867}, "spread_2dp": {"darwin": 0.0, "einstein": 0.019999999999999997, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.029999999999999916, "socrates": 0.019999999999999997}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.008531779010949772, "einstein": 0.011844572850719527, "hypatia": 0.006411547442246085, "ibn_sina": 0.0055551261976692105, "sagan": 0.9523236109174273, "socrates": 0.015333363580988076}, {"ibn_sina": 0.006040415676272804, "hypatia": 0.007690538369386058, "sagan": 0.9197382415954725, "socrates": 0.025252598393701732, "darwin": 0.00803448366844059, "einstein": 0.03324372229672636}, {"hypatia": 0.007134996644355875, "darwin": 0.007773296580586891, "sagan": 0.9285400990416, "ibn_sina": 0.006985232874675153, "einstein": 0.01665551910721172, "socrates": 0.032910855751570336}, {"ibn_sina": 0.006171692293940203, "einstein": 0.021713894524743396, "darwin": 0.009272974187968455, "sagan": 0.9232261517726976, "hypatia": 0.00813710458424046, "socrates": 0.031478182636409995}, {"einstein": 0.013406971592481278, "sagan": 0.9531598038069317, "socrates": 0.01312720962178747, "ibn_sina": 0.004799554677028032, "darwin": 0.005547152789777839, "hypatia": 0.00995930751199375}, {"socrates": 0.018782425577235407, "ibn_sina": 0.0037171857081694133, "hypatia": 0.006592846503460054, "einstein": 0.014674910205610273, "sagan": 0.9471154558324464, "darwin": 0.009117176173078603}], "prompt_tokens": 585, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.008531779010949772, "einstein": 0.011844572850719527, "hypatia": 0.006411547442246085, "ibn_sina": 0.0055551261976692105, "sagan": 0.9523236109174273, "socrates": 0.015333363580988076}, "confidence": 0.8889285093715579}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.006040415676272804, "hypatia": 0.007690538369386058, "sagan": 0.9197382415954725, "socrates": 0.025252598393701732, "darwin": 0.00803448366844059, "einstein": 0.03324372229672636}, "confidence": 0.8173857469368144}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.007134996644355875, "darwin": 0.007773296580586891, "sagan": 0.9285400990416, "ibn_sina": 0.006985232874675153, "einstein": 0.01665551910721172, "socrates": 0.032910855751570336}, "confidence": 0.8364488464796486}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.006171692293940203, "einstein": 0.021713894524743396, "darwin": 0.009272974187968455, "sagan": 0.9232261517726976, "hypatia": 0.00813710458424046, "socrates": 0.031478182636409995}, "confidence": 0.8247990241859483}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.013406971592481278, "sagan": 0.9531598038069317, "socrates": 0.01312720962178747, "ibn_sina": 0.004799554677028032, "darwin": 0.005547152789777839, "hypatia": 0.00995930751199375}, "confidence": 0.8908224118579525}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.018782425577235407, "ibn_sina": 0.0037171857081694133, "hypatia": 0.006592846503460054, "einstein": 0.014674910205610273, "sagan": 0.9471154558324464, "darwin": 0.009117176173078603}, "confidence": 0.8772834702073457}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:51Z"}
{"id": "c-einstein-01-main-03", "uid": "c027", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "prompt_sha256_by_order": ["7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af"], "choice": "einstein", "probs": {"darwin": 0.012845184503282321, "einstein": 0.891197889781719, "hypatia": 0.01062047524804032, "ibn_sina": 0.010688550890698683, "sagan": 0.04928028712176896, "socrates": 0.025367612454490675}, "confidence": 0.7572373355096739, "seconds": 0.2208566613332247, "seconds_all_orders": 1.3251399679993483, "seconds_by_order": [0.22077346200057946, 0.22069983300025342, 0.22070898299898545, 0.22066337300020678, 0.221722704000058, 0.2205716129992652], "argmax_stable": true, "spread": {"darwin": 0.00940789391793137, "einstein": 0.051447829034963366, "hypatia": 0.008227781105047832, "ibn_sina": 0.014800128494213707, "sagan": 0.03528591934124997, "socrates": 0.012956930374195624}, "spread_2dp": {"darwin": 0.01, "einstein": 0.04999999999999993, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.04, "socrates": 0.009999999999999998}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.012845184503282321, "einstein": 0.891197889781719, "hypatia": 0.01062047524804032, "ibn_sina": 0.010688550890698683, "sagan": 0.04928028712176896, "socrates": 0.025367612454490675}, {"ibn_sina": 0.008558907621410318, "hypatia": 0.008092989091793873, "sagan": 0.013994367780518994, "socrates": 0.016360004608726192, "darwin": 0.017277287046616394, "einstein": 0.9357164438509342}, {"hypatia": 0.0073801298749848485, "darwin": 0.012998583975136573, "sagan": 0.017477064175043758, "ibn_sina": 0.01287225406199248, "einstein": 0.9267001583375627, "socrates": 0.02257180957527966}, {"ibn_sina": 0.010975137875973172, "einstein": 0.9246707111463777, "darwin": 0.010408810745918942, "sagan": 0.02198040389750472, "hypatia": 0.013884362774796289, "socrates": 0.018080573559429262}, {"einstein": 0.9426457188166824, "sagan": 0.020772148070227064, "socrates": 0.017783674403994563, "ibn_sina": 0.005272483910662463, "darwin": 0.007869393128685023, "hypatia": 0.005656581669748456}, {"socrates": 0.029316934982921816, "ibn_sina": 0.02007261240487617, "hypatia": 0.009016620901176942, "einstein": 0.8924378619459366, "sagan": 0.039954335337646724, "darwin": 0.009201634427441778}], "prompt_tokens": 593, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.012845184503282321, "einstein": 0.891197889781719, "hypatia": 0.01062047524804032, "ibn_sina": 0.010688550890698683, "sagan": 0.04928028712176896, "socrates": 0.025367612454490675}, "confidence": 0.7572373355096739}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.008558907621410318, "hypatia": 0.008092989091793873, "sagan": 0.013994367780518994, "socrates": 0.016360004608726192, "darwin": 0.017277287046616394, "einstein": 0.9357164438509342}, "confidence": 0.8517592136718586}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.0073801298749848485, "darwin": 0.012998583975136573, "sagan": 0.017477064175043758, "ibn_sina": 0.01287225406199248, "einstein": 0.9267001583375627, "socrates": 0.02257180957527966}, "confidence": 0.8319726906994143}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.010975137875973172, "einstein": 0.9246707111463777, "darwin": 0.010408810745918942, "sagan": 0.02198040389750472, "hypatia": 0.013884362774796289, "socrates": 0.018080573559429262}, "confidence": 0.8274970502438634}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.9426457188166824, "sagan": 0.020772148070227064, "socrates": 0.017783674403994563, "ibn_sina": 0.005272483910662463, "darwin": 0.007869393128685023, "hypatia": 0.005656581669748456}, "confidence": 0.8673404989182003}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.029316934982921816, "ibn_sina": 0.02007261240487617, "hypatia": 0.009016620901176942, "einstein": 0.8924378619459366, "sagan": 0.039954335337646724, "darwin": 0.009201634427441778}, "confidence": 0.7593640579852824}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:52Z"}
{"id": "c-einstein-01-main-03", "uid": "c028", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "prompt_sha256_by_order": ["046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3"], "choice": "einstein", "probs": {"darwin": 0.014489821343511709, "einstein": 0.9337158590546821, "hypatia": 0.0061503260622735815, "ibn_sina": 0.007545481650684008, "sagan": 0.030476943744579925, "socrates": 0.007621568144268568}, "confidence": 0.8477403442933493, "seconds": 0.2240026071661608, "seconds_all_orders": 1.3440156429969647, "seconds_by_order": [0.22359502899962536, 0.22342466000009154, 0.22440596299929894, 0.22430629299924476, 0.22407550400021137, 0.2242081939984928], "argmax_stable": true, "spread": {"darwin": 0.08095036072288025, "einstein": 0.09060825482148871, "hypatia": 0.009309223735305789, "ibn_sina": 0.01004417481635769, "sagan": 0.02229723063982388, "socrates": 0.004234786969434276}, "spread_2dp": {"darwin": 0.09000000000000001, "einstein": 0.08999999999999997, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.019999999999999997, "socrates": 0.0}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.014489821343511709, "einstein": 0.9337158590546821, "hypatia": 0.0061503260622735815, "ibn_sina": 0.007545481650684008, "sagan": 0.030476943744579925, "socrates": 0.007621568144268568}, {"ibn_sina": 0.012257044149571137, "hypatia": 0.009564073327920992, "sagan": 0.010135427097193932, "socrates": 0.006460780535683516, "darwin": 0.023254298542475738, "einstein": 0.9383283763471547}, {"hypatia": 0.006443640106093338, "darwin": 0.0951812334033792, "sagan": 0.020790681225608922, "ibn_sina": 0.010057463414010112, "einstein": 0.8574695184368983, "socrates": 0.010057463414010112}, {"ibn_sina": 0.010117931615648194, "einstein": 0.9401864999429483, "darwin": 0.01428207822248442, "sagan": 0.01265893006412209, "hypatia": 0.01476667887205722, "socrates": 0.007987881282739773}, {"einstein": 0.948077773258387, "sagan": 0.020372335548650526, "socrates": 0.005822676444575836, "ibn_sina": 0.00603888693113624, "darwin": 0.014230872680498942, "hypatia": 0.005457455136751431}, {"socrates": 0.00876168945531895, "ibn_sina": 0.01608306174749393, "hypatia": 0.00577426278824447, "einstein": 0.9112391379038376, "sagan": 0.03243265773701781, "darwin": 0.02570919036808731}], "prompt_tokens": 623, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.014489821343511709, "einstein": 0.9337158590546821, "hypatia": 0.0061503260622735815, "ibn_sina": 0.007545481650684008, "sagan": 0.030476943744579925, "socrates": 0.007621568144268568}, "confidence": 0.8477403442933493}}, "usage": {"input_tokens": 623, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.012257044149571137, "hypatia": 0.009564073327920992, "sagan": 0.010135427097193932, "socrates": 0.006460780535683516, "darwin": 0.023254298542475738, "einstein": 0.9383283763471547}, "confidence": 0.8576644953477632}}, "usage": {"input_tokens": 623, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.006443640106093338, "darwin": 0.0951812334033792, "sagan": 0.020790681225608922, "ibn_sina": 0.010057463414010112, "einstein": 0.8574695184368983, "socrates": 0.010057463414010112}, "confidence": 0.6939874243658618}}, "usage": {"input_tokens": 623, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.010117931615648194, "einstein": 0.9401864999429483, "darwin": 0.01428207822248442, "sagan": 0.01265893006412209, "hypatia": 0.01476667887205722, "socrates": 0.007987881282739773}, "confidence": 0.8616389374434036}}, "usage": {"input_tokens": 623, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.948077773258387, "sagan": 0.020372335548650526, "socrates": 0.005822676444575836, "ibn_sina": 0.00603888693113624, "darwin": 0.014230872680498942, "hypatia": 0.005457455136751431}, "confidence": 0.8794830033669424}}, "usage": {"input_tokens": 623, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.00876168945531895, "ibn_sina": 0.01608306174749393, "hypatia": 0.00577426278824447, "einstein": 0.9112391379038376, "sagan": 0.03243265773701781, "darwin": 0.02570919036808731}, "confidence": 0.7989260564716373}}, "usage": {"input_tokens": 623, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:54Z"}
{"id": "c-einstein-01-main-03", "uid": "c029", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "prompt_sha256_by_order": ["1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979"], "choice": "einstein", "probs": {"darwin": 0.006095348652413989, "einstein": 0.9552798992126481, "hypatia": 0.008405735690598503, "ibn_sina": 0.004862714888651956, "sagan": 0.01935530189792627, "socrates": 0.006000999657761125}, "confidence": 0.895722137454355, "seconds": 0.25478004049970576, "seconds_all_orders": 1.5286802429982345, "seconds_by_order": [0.2552588580001611, 0.25470339299863554, 0.25470300300003146, 0.2545402239993564, 0.25443314500080305, 0.2550416199992469], "argmax_stable": true, "spread": {"darwin": 0.0050844267739558885, "einstein": 0.01733170201343892, "hypatia": 0.006492487416200529, "ibn_sina": 0.006026985449357727, "sagan": 0.013373633225312702, "socrates": 0.00448333057968083}, "spread_2dp": {"darwin": 0.01, "einstein": 0.020000000000000018, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.006095348652413989, "einstein": 0.9552798992126481, "hypatia": 0.008405735690598503, "ibn_sina": 0.004862714888651956, "sagan": 0.01935530189792627, "socrates": 0.006000999657761125}, {"ibn_sina": 0.004495229799730816, "hypatia": 0.008115947638400234, "sagan": 0.0067433146253170605, "socrates": 0.005903734008106275, "darwin": 0.009359068155596551, "einstein": 0.9653827057728491}, {"hypatia": 0.006563136482052402, "darwin": 0.006190579752765303, "sagan": 0.005981668672613568, "ibn_sina": 0.007952472893870247, "einstein": 0.9639551162970283, "socrates": 0.009357025901670046}, {"ibn_sina": 0.007780475593860052, "einstein": 0.9590588798594133, "darwin": 0.00722171799183983, "sagan": 0.008186268278111645, "hypatia": 0.011239243659422872, "socrates": 0.006513414617352267}, {"einstein": 0.9718823350005974, "sagan": 0.01093610295631032, "socrates": 0.004873695321989216, "ibn_sina": 0.003286469096239985, "darwin": 0.004274641381640663, "hypatia": 0.004746756243222343}, {"socrates": 0.008269044624605747, "ibn_sina": 0.009313454545597712, "hypatia": 0.007225598775235949, "einstein": 0.9545506329871585, "sagan": 0.01553346013273012, "darwin": 0.005107808934671999}], "prompt_tokens": 707, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.006095348652413989, "einstein": 0.9552798992126481, "hypatia": 0.008405735690598503, "ibn_sina": 0.004862714888651956, "sagan": 0.01935530189792627, "socrates": 0.006000999657761125}, "confidence": 0.895722137454355}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.004495229799730816, "hypatia": 0.008115947638400234, "sagan": 0.0067433146253170605, "socrates": 0.005903734008106275, "darwin": 0.009359068155596551, "einstein": 0.9653827057728491}, "confidence": 0.9186613153917336}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.006563136482052402, "darwin": 0.006190579752765303, "sagan": 0.005981668672613568, "ibn_sina": 0.007952472893870247, "einstein": 0.9639551162970283, "socrates": 0.009357025901670046}, "confidence": 0.9153729284708032}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.007780475593860052, "einstein": 0.9590588798594133, "darwin": 0.00722171799183983, "sagan": 0.008186268278111645, "hypatia": 0.011239243659422872, "socrates": 0.006513414617352267}, "confidence": 0.904170861045697}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.9718823350005974, "sagan": 0.01093610295631032, "socrates": 0.004873695321989216, "ibn_sina": 0.003286469096239985, "darwin": 0.004274641381640663, "hypatia": 0.004746756243222343}, "confidence": 0.9337002753676849}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.008269044624605747, "ibn_sina": 0.009313454545597712, "hypatia": 0.007225598775235949, "einstein": 0.9545506329871585, "sagan": 0.01553346013273012, "darwin": 0.005107808934671999}, "confidence": 0.8939699390130766}}, "usage": {"input_tokens": 707, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:55Z"}
{"id": "c-einstein-01-starter-01", "uid": "c030", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "prompt_sha256_by_order": ["08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa"], "choice": "einstein", "probs": {"darwin": 0.012063235364831943, "einstein": 0.964090130627568, "hypatia": 0.005519019393262553, "ibn_sina": 0.004312622808934852, "sagan": 0.007540029859555299, "socrates": 0.006474961945847441}, "confidence": 0.9157157645127842, "seconds": 0.18677483699987837, "seconds_all_orders": 1.1206490219992702, "seconds_by_order": [0.18644747500002268, 0.18662344399854192, 0.18677852200016787, 0.18705002900060208, 0.18747009599974263, 0.18627945600019302], "argmax_stable": true, "spread": {"darwin": 0.006714216021586401, "einstein": 0.021815771499257508, "hypatia": 0.0024116224269040577, "ibn_sina": 0.007061017957829023, "sagan": 0.008890218275639652, "socrates": 0.007028191229713109}, "spread_2dp": {"darwin": 0.01, "einstein": 0.020000000000000018, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.012063235364831943, "einstein": 0.964090130627568, "hypatia": 0.005519019393262553, "ibn_sina": 0.004312622808934852, "sagan": 0.007540029859555299, "socrates": 0.006474961945847441}, {"ibn_sina": 0.003996871604028714, "hypatia": 0.005777553093026303, "sagan": 0.005041994839629114, "socrates": 0.004909846019436001, "darwin": 0.018438173166255095, "einstein": 0.9618355612776248}, {"hypatia": 0.005802028351330816, "darwin": 0.018136617250797953, "sagan": 0.006879831614246016, "ibn_sina": 0.007261077684986964, "einstein": 0.9514154085980311, "socrates": 0.010505036500607263}, {"ibn_sina": 0.008181785680148221, "einstein": 0.958605215194871, "darwin": 0.01450858534893806, "sagan": 0.005831197993558607, "hypatia": 0.007299897515238219, "socrates": 0.005573318267245877}, {"einstein": 0.9629982203323151, "sagan": 0.012341796715420007, "socrates": 0.004611560570413087, "ibn_sina": 0.0034361901488488165, "darwin": 0.011723957144668694, "hypatia": 0.004888275088334161}, {"socrates": 0.011639751800126196, "ibn_sina": 0.01049720810667784, "hypatia": 0.005635708318242942, "einstein": 0.9422743591283105, "sagan": 0.013932213115268766, "darwin": 0.016020759531373833}], "prompt_tokens": 515, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.012063235364831943, "einstein": 0.964090130627568, "hypatia": 0.005519019393262553, "ibn_sina": 0.004312622808934852, "sagan": 0.007540029859555299, "socrates": 0.006474961945847441}, "confidence": 0.9157157645127842}}, "usage": {"input_tokens": 515, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.003996871604028714, "hypatia": 0.005777553093026303, "sagan": 0.005041994839629114, "socrates": 0.004909846019436001, "darwin": 0.018438173166255095, "einstein": 0.9618355612776248}, "confidence": 0.910679795884254}}, "usage": {"input_tokens": 515, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.005802028351330816, "darwin": 0.018136617250797953, "sagan": 0.006879831614246016, "ibn_sina": 0.007261077684986964, "einstein": 0.9514154085980311, "socrates": 0.010505036500607263}, "confidence": 0.8869171495121403}}, "usage": {"input_tokens": 515, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.008181785680148221, "einstein": 0.958605215194871, "darwin": 0.01450858534893806, "sagan": 0.005831197993558607, "hypatia": 0.007299897515238219, "socrates": 0.005573318267245877}, "confidence": 0.903183703017795}}, "usage": {"input_tokens": 515, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.9629982203323151, "sagan": 0.012341796715420007, "socrates": 0.004611560570413087, "ibn_sina": 0.0034361901488488165, "darwin": 0.011723957144668694, "hypatia": 0.004888275088334161}, "confidence": 0.9132547751289684}}, "usage": {"input_tokens": 515, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.011639751800126196, "ibn_sina": 0.01049720810667784, "hypatia": 0.005635708318242942, "einstein": 0.9422743591283105, "sagan": 0.013932213115268766, "darwin": 0.016020759531373833}, "confidence": 0.8663310106926059}}, "usage": {"input_tokens": 515, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:56Z"}
{"id": "c-einstein-01-starter-02", "uid": "c031", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "prompt_sha256_by_order": ["06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7"], "choice": "hypatia", "probs": {"darwin": 0.0179015089400173, "einstein": 0.38583707350417384, "hypatia": 0.4147550366076549, "ibn_sina": 0.11769784649606058, "sagan": 0.01608128949878332, "socrates": 0.047727244953309995}, "confidence": 0.20512207906791885, "seconds": 0.18728673250037295, "seconds_all_orders": 1.1237203950022376, "seconds_by_order": [0.18713362900052743, 0.18703879900022002, 0.18715148900082568, 0.18787949299985485, 0.18754228500074532, 0.18697470000006433], "argmax_stable": false, "spread": {"darwin": 0.005533083723010073, "einstein": 0.14011756133998116, "hypatia": 0.1284106482032466, "ibn_sina": 0.06611273377582817, "sagan": 0.00832920604875734, "socrates": 0.032706326105124105}, "spread_2dp": {"darwin": 0.01, "einstein": 0.14, "hypatia": 0.12, "ibn_sina": 0.06, "sagan": 0.0, "socrates": 0.030000000000000002}, "choices_by_order": ["hypatia", "einstein", "hypatia", "hypatia", "einstein", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0179015089400173, "einstein": 0.38583707350417384, "hypatia": 0.4147550366076549, "ibn_sina": 0.11769784649606058, "sagan": 0.01608128949878332, "socrates": 0.047727244953309995}, {"ibn_sina": 0.1268362202328298, "hypatia": 0.3260410646974154, "sagan": 0.016960991124083804, "socrates": 0.017067522811955053, "darwin": 0.016960991124083804, "einstein": 0.4961332100096321}, {"hypatia": 0.454451712900662, "darwin": 0.013404212921137908, "sagan": 0.02029964499448267, "ibn_sina": 0.10174867910574337, "einstein": 0.3768125682094825, "socrates": 0.033283181868491524}, {"ibn_sina": 0.13146286470210206, "einstein": 0.370438594013495, "darwin": 0.012368425217007227, "sagan": 0.023230079693993357, "hypatia": 0.42647928997949897, "socrates": 0.036020746393903444}, {"einstein": 0.4647823218881325, "sagan": 0.02441049554754066, "socrates": 0.01502091884818589, "ibn_sina": 0.06535013092627388, "darwin": 0.016011211432755208, "hypatia": 0.41442492135711184}, {"socrates": 0.033417970372048585, "ibn_sina": 0.12349726499917969, "hypatia": 0.4492021721345667, "einstein": 0.35601564866965096, "sagan": 0.021088020511043686, "darwin": 0.016778923313510344}], "prompt_tokens": 441, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"darwin": 0.0179015089400173, "einstein": 0.38583707350417384, "hypatia": 0.4147550366076549, "ibn_sina": 0.11769784649606058, "sagan": 0.01608128949878332, "socrates": 0.047727244953309995}, "confidence": 0.20512207906791885}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.1268362202328298, "hypatia": 0.3260410646974154, "sagan": 0.016960991124083804, "socrates": 0.017067522811955053, "darwin": 0.016960991124083804, "einstein": 0.4961332100096321}, "confidence": 0.24328601857743512}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.454451712900662, "darwin": 0.013404212921137908, "sagan": 0.02029964499448267, "ibn_sina": 0.10174867910574337, "einstein": 0.3768125682094825, "socrates": 0.033283181868491524}, "confidence": 0.2326796599899561}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.13146286470210206, "einstein": 0.370438594013495, "darwin": 0.012368425217007227, "sagan": 0.023230079693993357, "hypatia": 0.42647928997949897, "socrates": 0.036020746393903444}, "confidence": 0.2060583162730092}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.4647823218881325, "sagan": 0.02441049554754066, "socrates": 0.01502091884818589, "ibn_sina": 0.06535013092627388, "darwin": 0.016011211432755208, "hypatia": 0.41442492135711184}, "confidence": 0.2717429451773413}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"socrates": 0.033417970372048585, "ibn_sina": 0.12349726499917969, "hypatia": 0.4492021721345667, "einstein": 0.35601564866965096, "sagan": 0.021088020511043686, "darwin": 0.016778923313510344}, "confidence": 0.2147491667569418}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:57Z"}
{"id": "c-einstein-01-starter-02", "uid": "c032", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "prompt_sha256_by_order": ["8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db"], "choice": "sagan", "probs": {"darwin": 0.05373741553771287, "einstein": 0.021094030711322726, "hypatia": 0.012458857518454186, "ibn_sina": 0.013873082865346827, "sagan": 0.8802621640858854, "socrates": 0.01857444928127793}, "confidence": 0.7346642094477354, "seconds": 0.19231808116637694, "seconds_all_orders": 1.1539084869982617, "seconds_by_order": [0.1918849499998032, 0.1922353669997392, 0.19285087199932605, 0.19261153399929754, 0.1921794569989288, 0.19214630700116686], "argmax_stable": true, "spread": {"darwin": 0.055123329086945935, "einstein": 0.0218807551578524, "hypatia": 0.0037425173795695415, "ibn_sina": 0.006815187159358226, "sagan": 0.06854416006428765, "socrates": 0.0040387021525055515}, "spread_2dp": {"darwin": 0.06, "einstein": 0.02, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.07000000000000006, "socrates": 0.0}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.05373741553771287, "einstein": 0.021094030711322726, "hypatia": 0.012458857518454186, "ibn_sina": 0.013873082865346827, "sagan": 0.8802621640858854, "socrates": 0.01857444928127793}, {"ibn_sina": 0.008022996425343192, "hypatia": 0.010239878509699452, "sagan": 0.833697598550443, "socrates": 0.016785641363151443, "darwin": 0.08838839430656595, "einstein": 0.04286549084479695}, {"hypatia": 0.010511445607451524, "darwin": 0.05938078214191979, "sagan": 0.8776996982033628, "ibn_sina": 0.010916406592627756, "einstein": 0.020984735686944547, "socrates": 0.020506931767693602}, {"ibn_sina": 0.008579436059601563, "einstein": 0.023297106216628314, "darwin": 0.051634934226751715, "sagan": 0.8888878643330711, "hypatia": 0.011132429548759188, "socrates": 0.01646822961518805}, {"einstein": 0.024062305154880496, "sagan": 0.8685984895837575, "socrates": 0.017458113736650575, "ibn_sina": 0.013396753775568291, "darwin": 0.06250194185987405, "hypatia": 0.013982395889268993}, {"socrates": 0.01647796149886005, "ibn_sina": 0.007057895705988602, "hypatia": 0.011664171128048078, "einstein": 0.02929314783275258, "sagan": 0.9022417586147307, "darwin": 0.03326506521962001}], "prompt_tokens": 495, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.05373741553771287, "einstein": 0.021094030711322726, "hypatia": 0.012458857518454186, "ibn_sina": 0.013873082865346827, "sagan": 0.8802621640858854, "socrates": 0.01857444928127793}, "confidence": 0.7346642094477354}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.008022996425343192, "hypatia": 0.010239878509699452, "sagan": 0.833697598550443, "socrates": 0.016785641363151443, "darwin": 0.08838839430656595, "einstein": 0.04286549084479695}, "confidence": 0.6461831508660889}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.010511445607451524, "darwin": 0.05938078214191979, "sagan": 0.8776996982033628, "ibn_sina": 0.010916406592627756, "einstein": 0.020984735686944547, "socrates": 0.020506931767693602}, "confidence": 0.729968067181708}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.008579436059601563, "einstein": 0.023297106216628314, "darwin": 0.051634934226751715, "sagan": 0.8888878643330711, "hypatia": 0.011132429548759188, "socrates": 0.01646822961518805}, "confidence": 0.7525591566960346}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.024062305154880496, "sagan": 0.8685984895837575, "socrates": 0.017458113736650575, "ibn_sina": 0.013396753775568291, "darwin": 0.06250194185987405, "hypatia": 0.013982395889268993}, "confidence": 0.711554307417491}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.01647796149886005, "ibn_sina": 0.007057895705988602, "hypatia": 0.011664171128048078, "einstein": 0.02929314783275258, "sagan": 0.9022417586147307, "darwin": 0.03326506521962001}, "confidence": 0.7797546808687335}}, "usage": {"input_tokens": 495, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:17:59Z"}
{"id": "c-einstein-01-starter-03", "uid": "c033", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "prompt_sha256_by_order": ["ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c"], "choice": "einstein", "probs": {"darwin": 0.014239444625543399, "einstein": 0.8040552027389014, "hypatia": 0.047842817138422505, "ibn_sina": 0.015698728921121646, "sagan": 0.044054436019833845, "socrates": 0.07410937055617725}, "confidence": 0.5880110898561732, "seconds": 0.2303905210001176, "seconds_all_orders": 1.3823431260007055, "seconds_by_order": [0.23040073300035147, 0.23061485099970014, 0.23032825300106197, 0.2300811650002288, 0.23018555399903562, 0.23073257000032754], "argmax_stable": true, "spread": {"darwin": 0.010371056484932424, "einstein": 0.07098551021790933, "hypatia": 0.1422896064712194, "ibn_sina": 0.04521729172677953, "sagan": 0.06339963143391401, "socrates": 0.06511228018074552}, "spread_2dp": {"darwin": 0.01, "einstein": 0.07000000000000006, "hypatia": 0.14, "ibn_sina": 0.04, "sagan": 0.07, "socrates": 0.06999999999999999}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.014239444625543399, "einstein": 0.8040552027389014, "hypatia": 0.047842817138422505, "ibn_sina": 0.015698728921121646, "sagan": 0.044054436019833845, "socrates": 0.07410937055617725}, {"ibn_sina": 0.05323721930217305, "hypatia": 0.09135567065288186, "sagan": 0.0404016702390682, "socrates": 0.02385627875340159, "darwin": 0.01895315790626244, "einstein": 0.772196003146213}, {"hypatia": 0.1814982956482666, "darwin": 0.010130299969578463, "sagan": 0.014104663339032026, "ibn_sina": 0.012448773429243934, "einstein": 0.7404898332153163, "socrates": 0.04132813439856267}, {"ibn_sina": 0.028456438231491004, "einstein": 0.8001580890343836, "darwin": 0.011351901889591797, "sagan": 0.014513043494954749, "hypatia": 0.05735549281846006, "socrates": 0.08816503453111879}, {"einstein": 0.7330696925209921, "sagan": 0.03714310052226921, "socrates": 0.023052754350373274, "ibn_sina": 0.008019927575393517, "darwin": 0.008582101421330014, "hypatia": 0.19013242360964192}, {"socrates": 0.055291033141981236, "ibn_sina": 0.0338025147732276, "hypatia": 0.049604405480049664, "einstein": 0.7712099573016673, "sagan": 0.07750429477294604, "darwin": 0.012587794530128267}], "prompt_tokens": 677, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.014239444625543399, "einstein": 0.8040552027389014, "hypatia": 0.047842817138422505, "ibn_sina": 0.015698728921121646, "sagan": 0.044054436019833845, "socrates": 0.07410937055617725}, "confidence": 0.5880110898561732}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.05323721930217305, "hypatia": 0.09135567065288186, "sagan": 0.0404016702390682, "socrates": 0.02385627875340159, "darwin": 0.01895315790626244, "einstein": 0.772196003146213}, "confidence": 0.5320328398517653}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.1814982956482666, "darwin": 0.010130299969578463, "sagan": 0.014104663339032026, "ibn_sina": 0.012448773429243934, "einstein": 0.7404898332153163, "socrates": 0.04132813439856267}, "confidence": 0.5001176506919343}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.028456438231491004, "einstein": 0.8001580890343836, "darwin": 0.011351901889591797, "sagan": 0.014513043494954749, "hypatia": 0.05735549281846006, "socrates": 0.08816503453111879}, "confidence": 0.5829579475627581}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.7330696925209921, "sagan": 0.03714310052226921, "socrates": 0.023052754350373274, "ibn_sina": 0.008019927575393517, "darwin": 0.008582101421330014, "hypatia": 0.19013242360964192}, "confidence": 0.4907086284437977}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.055291033141981236, "ibn_sina": 0.0338025147732276, "hypatia": 0.049604405480049664, "einstein": 0.7712099573016673, "sagan": 0.07750429477294604, "darwin": 0.012587794530128267}, "confidence": 0.529108566297466}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:00Z"}
{"id": "c-einstein-01-starter-03", "uid": "c034", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "prompt_sha256_by_order": ["70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c"], "choice": "einstein", "probs": {"darwin": 0.005852802257439867, "einstein": 0.9718068266567751, "hypatia": 0.005374850631836261, "ibn_sina": 0.00486739551636046, "sagan": 0.004786169436480454, "socrates": 0.007311955501107874}, "confidence": 0.9334860595606338, "seconds": 0.22177828849968742, "seconds_all_orders": 1.3306697309981246, "seconds_by_order": [0.2221714700008306, 0.2209944899987022, 0.22135858699948585, 0.22210683099910966, 0.22202467100032663, 0.22201368199966964], "argmax_stable": true, "spread": {"darwin": 0.0027577363367276434, "einstein": 0.009252806072021502, "hypatia": 0.002584522674118464, "ibn_sina": 0.0026475546141503953, "sagan": 0.003056823407919752, "socrates": 0.004047883117885704}, "spread_2dp": {"darwin": 0.01, "einstein": 0.010000000000000009, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.005852802257439867, "einstein": 0.9718068266567751, "hypatia": 0.005374850631836261, "ibn_sina": 0.00486739551636046, "sagan": 0.004786169436480454, "socrates": 0.007311955501107874}, {"ibn_sina": 0.005898124231285804, "hypatia": 0.005815260257823441, "sagan": 0.003157686927705695, "socrates": 0.004766476972027864, "darwin": 0.006841861897445935, "einstein": 0.9735205897137112}, {"hypatia": 0.003913701843461244, "darwin": 0.005888313890753598, "sagan": 0.003531422962646318, "ibn_sina": 0.005280082324852891, "einstein": 0.9735255458301108, "socrates": 0.007860933148175216}, {"ibn_sina": 0.006471494727787977, "einstein": 0.9701658850895838, "darwin": 0.005345708108962302, "sagan": 0.0035278996557388934, "hypatia": 0.006498224517579708, "socrates": 0.007990787900347412}, {"einstein": 0.9773457101392133, "sagan": 0.005575519734627601, "socrates": 0.004804288347859334, "ibn_sina": 0.0038239401136375816, "darwin": 0.004084125560718291, "hypatia": 0.004366416103943766}, {"socrates": 0.008814360089913568, "ibn_sina": 0.0063747959458556915, "hypatia": 0.005045917723385467, "einstein": 0.9680929040671918, "sagan": 0.006214510335625447, "darwin": 0.005457511838028006}], "prompt_tokens": 593, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.005852802257439867, "einstein": 0.9718068266567751, "hypatia": 0.005374850631836261, "ibn_sina": 0.00486739551636046, "sagan": 0.004786169436480454, "socrates": 0.007311955501107874}, "confidence": 0.9334860595606338}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.005898124231285804, "hypatia": 0.005815260257823441, "sagan": 0.003157686927705695, "socrates": 0.004766476972027864, "darwin": 0.006841861897445935, "einstein": 0.9735205897137112}, "confidence": 0.9374685340978335}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.003913701843461244, "darwin": 0.005888313890753598, "sagan": 0.003531422962646318, "ibn_sina": 0.005280082324852891, "einstein": 0.9735255458301108, "socrates": 0.007860933148175216}, "confidence": 0.9374849466086428}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.006471494727787977, "einstein": 0.9701658850895838, "darwin": 0.005345708108962302, "sagan": 0.0035278996557388934, "hypatia": 0.006498224517579708, "socrates": 0.007990787900347412}, "confidence": 0.9296929925439805}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.9773457101392133, "sagan": 0.005575519734627601, "socrates": 0.004804288347859334, "ibn_sina": 0.0038239401136375816, "darwin": 0.004084125560718291, "hypatia": 0.004366416103943766}, "confidence": 0.946371007508254}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.008814360089913568, "ibn_sina": 0.0063747959458556915, "hypatia": 0.005045917723385467, "einstein": 0.9680929040671918, "sagan": 0.006214510335625447, "darwin": 0.005457511838028006}, "confidence": 0.924899281278688}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:01Z"}
{"id": "c-einstein-01-starter-03", "uid": "c035", "task": "kperm", "label": "einstein", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "prompt_sha256_by_order": ["36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156"], "choice": "einstein", "probs": {"darwin": 0.011440127840955492, "einstein": 0.9308791245625871, "hypatia": 0.023810167857353612, "ibn_sina": 0.011544199000632848, "sagan": 0.013851309172243677, "socrates": 0.008475071566227364}, "confidence": 0.8411568391590704, "seconds": 0.18937823366619946, "seconds_all_orders": 1.1362694019971968, "seconds_by_order": [0.1888424849985313, 0.1894135410002491, 0.1901480439992156, 0.18996501499896112, 0.1891571220003243, 0.18874319499991543], "argmax_stable": true, "spread": {"darwin": 0.003201754817169118, "einstein": 0.013644143363944483, "hypatia": 0.012074248824634218, "ibn_sina": 0.005243940658712915, "sagan": 0.006162492815706506, "socrates": 0.007031395237719106}, "spread_2dp": {"darwin": 0.0, "einstein": 0.010000000000000009, "hypatia": 0.02, "ibn_sina": 0.0, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.011440127840955492, "einstein": 0.9308791245625871, "hypatia": 0.023810167857353612, "ibn_sina": 0.011544199000632848, "sagan": 0.013851309172243677, "socrates": 0.008475071566227364}, {"ibn_sina": 0.009379259748109096, "hypatia": 0.029576262813840584, "sagan": 0.01028840940011567, "socrates": 0.011390240741591418, "darwin": 0.010953156575477116, "einstein": 0.9284126707208662}, {"hypatia": 0.03588441668198783, "darwin": 0.008238373023786374, "sagan": 0.010853814436671732, "ibn_sina": 0.012036056171912928, "einstein": 0.9174808728816948, "socrates": 0.01550646680394647}, {"ibn_sina": 0.012859959456041573, "einstein": 0.9292554384071867, "darwin": 0.009946825844801605, "sagan": 0.013135942191929411, "hypatia": 0.025866915441126313, "socrates": 0.008934918658914364}, {"einstein": 0.9311250162456393, "sagan": 0.012614196434902608, "socrates": 0.009948500837712433, "ibn_sina": 0.00836610804272517, "darwin": 0.008746466120213459, "hypatia": 0.029199712318807004}, {"socrates": 0.012972600772446202, "ibn_sina": 0.013610048701438084, "hypatia": 0.0240412706894979, "einstein": 0.9240376336664672, "sagan": 0.016450902215822176, "darwin": 0.00888754395432846}], "prompt_tokens": 463, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.011440127840955492, "einstein": 0.9308791245625871, "hypatia": 0.023810167857353612, "ibn_sina": 0.011544199000632848, "sagan": 0.013851309172243677, "socrates": 0.008475071566227364}, "confidence": 0.8411568391590704}}, "usage": {"input_tokens": 463, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.009379259748109096, "hypatia": 0.029576262813840584, "sagan": 0.01028840940011567, "socrates": 0.011390240741591418, "darwin": 0.010953156575477116, "einstein": 0.9284126707208662}, "confidence": 0.8359220482979322}}, "usage": {"input_tokens": 463, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.03588441668198783, "darwin": 0.008238373023786374, "sagan": 0.010853814436671732, "ibn_sina": 0.012036056171912928, "einstein": 0.9174808728816948, "socrates": 0.01550646680394647}, "confidence": 0.8123558040438198}}, "usage": {"input_tokens": 463, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.012859959456041573, "einstein": 0.9292554384071867, "darwin": 0.009946825844801605, "sagan": 0.013135942191929411, "hypatia": 0.025866915441126313, "socrates": 0.008934918658914364}, "confidence": 0.8376417649288962}}, "usage": {"input_tokens": 463, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.9311250162456393, "sagan": 0.012614196434902608, "socrates": 0.009948500837712433, "ibn_sina": 0.00836610804272517, "darwin": 0.008746466120213459, "hypatia": 0.029199712318807004}, "confidence": 0.8419012025583207}}, "usage": {"input_tokens": 463, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.012972600772446202, "ibn_sina": 0.013610048701438084, "hypatia": 0.0240412706894979, "einstein": 0.9240376336664672, "sagan": 0.016450902215822176, "darwin": 0.00888754395432846}, "confidence": 0.8261520082552016}}, "usage": {"input_tokens": 463, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:02Z"}
{"id": "c-hypatia-01-dessert-01", "uid": "c036", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "prompt_sha256_by_order": ["53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816"], "choice": "einstein", "probs": {"darwin": 0.10362642172448538, "einstein": 0.3477044148491018, "hypatia": 0.17496399630990866, "ibn_sina": 0.20547382338929432, "sagan": 0.13319093732981493, "socrates": 0.035040406397394905}, "confidence": 0.06812361202625813, "seconds": 0.2535249088335452, "seconds_all_orders": 1.5211494530012715, "seconds_by_order": [0.2534344529994996, 0.2539972780014068, 0.2539454090001527, 0.2521929029990133, 0.2535799520010187, 0.2539994580001803], "argmax_stable": false, "spread": {"darwin": 0.05580321174599151, "einstein": 0.25299994391444236, "hypatia": 0.12533542890106372, "ibn_sina": 0.09270721599985099, "sagan": 0.30891302530332765, "socrates": 0.016881961290761027}, "spread_2dp": {"darwin": 0.06, "einstein": 0.25, "hypatia": 0.12000000000000001, "ibn_sina": 0.09999999999999999, "sagan": 0.31, "socrates": 0.010000000000000002}, "choices_by_order": ["einstein", "sagan", "einstein", "einstein", "hypatia", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.10362642172448538, "einstein": 0.3477044148491018, "hypatia": 0.17496399630990866, "ibn_sina": 0.20547382338929432, "sagan": 0.13319093732981493, "socrates": 0.035040406397394905}, {"ibn_sina": 0.11276660738944333, "hypatia": 0.10783288159086873, "sagan": 0.44210396263314256, "socrates": 0.04049565519580908, "darwin": 0.1594296334704769, "einstein": 0.13737125972025943}, {"hypatia": 0.2324038219274191, "darwin": 0.1056573256383376, "sagan": 0.19358866506014172, "ibn_sina": 0.1359343133325876, "einstein": 0.280493506353358, "socrates": 0.05192236768815593}, {"ibn_sina": 0.1299543295682922, "einstein": 0.3903712036347018, "darwin": 0.11641052677600436, "sagan": 0.14652852203616046, "hypatia": 0.1720375579858587, "socrates": 0.04469785999898248}, {"einstein": 0.2288499180760071, "sagan": 0.2234709499047715, "socrates": 0.04684650003818991, "ibn_sina": 0.14790681165629044, "darwin": 0.11975750983280867, "hypatia": 0.23316831049193246}, {"socrates": 0.03575997763115213, "ibn_sina": 0.16953924972996487, "hypatia": 0.13675922154547612, "einstein": 0.36186403547209084, "sagan": 0.16382080938503976, "darwin": 0.1322567062362762}], "prompt_tokens": 693, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.10362642172448538, "einstein": 0.3477044148491018, "hypatia": 0.17496399630990866, "ibn_sina": 0.20547382338929432, "sagan": 0.13319093732981493, "socrates": 0.035040406397394905}, "confidence": 0.06812361202625813}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.11276660738944333, "hypatia": 0.10783288159086873, "sagan": 0.44210396263314256, "socrates": 0.04049565519580908, "darwin": 0.1594296334704769, "einstein": 0.13737125972025943}, "confidence": 0.11887446518277005}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.2324038219274191, "darwin": 0.1056573256383376, "sagan": 0.19358866506014172, "ibn_sina": 0.1359343133325876, "einstein": 0.280493506353358, "socrates": 0.05192236768815593}, "confidence": 0.04300270607347891}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.1299543295682922, "einstein": 0.3903712036347018, "darwin": 0.11641052677600436, "sagan": 0.14652852203616046, "hypatia": 0.1720375579858587, "socrates": 0.04469785999898248}, "confidence": 0.0830735715536138}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.2288499180760071, "sagan": 0.2234709499047715, "socrates": 0.04684650003818991, "ibn_sina": 0.14790681165629044, "darwin": 0.11975750983280867, "hypatia": 0.23316831049193246}, "confidence": 0.034110270560693226}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.03575997763115213, "ibn_sina": 0.16953924972996487, "hypatia": 0.13675922154547612, "einstein": 0.36186403547209084, "sagan": 0.16382080938503976, "darwin": 0.1322567062362762}, "confidence": 0.06880011037250719}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:04Z"}
{"id": "c-hypatia-01-dessert-02", "uid": "c037", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "prompt_sha256_by_order": ["60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b"], "choice": "socrates", "probs": {"darwin": 0.04599934472801319, "einstein": 0.11097956656593166, "hypatia": 0.2865316813630734, "ibn_sina": 0.09343128662340078, "sagan": 0.11288636275408764, "socrates": 0.3501717579654933}, "confidence": 0.08875096555349686, "seconds": 0.19114431433354184, "seconds_all_orders": 1.146865886001251, "seconds_by_order": [0.19140295400029572, 0.19064883000100963, 0.19029252300060762, 0.1917514110009506, 0.19118566599900078, 0.1915845019993867], "argmax_stable": false, "spread": {"darwin": 0.010840604221185483, "einstein": 0.12939631446985883, "hypatia": 0.11086661836002573, "ibn_sina": 0.08452361056312824, "sagan": 0.014098896736956579, "socrates": 0.15977995011341428}, "spread_2dp": {"darwin": 0.010000000000000002, "einstein": 0.13, "hypatia": 0.10999999999999999, "ibn_sina": 0.09, "sagan": 0.009999999999999995, "socrates": 0.16}, "choices_by_order": ["socrates", "hypatia", "socrates", "socrates", "hypatia", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.04599934472801319, "einstein": 0.11097956656593166, "hypatia": 0.2865316813630734, "ibn_sina": 0.09343128662340078, "sagan": 0.11288636275408764, "socrates": 0.3501717579654933}, {"ibn_sina": 0.15757634831267028, "hypatia": 0.3145041751457066, "sagan": 0.10632937716209864, "socrates": 0.2543279025451256, "darwin": 0.045411893242706496, "einstein": 0.12185030359169231}, {"hypatia": 0.23815634459996815, "darwin": 0.035158740506827706, "sagan": 0.12042827389905522, "ibn_sina": 0.07305273774954205, "einstein": 0.1623491375786141, "socrates": 0.3708547656659928}, {"ibn_sina": 0.14253967994861494, "einstein": 0.10325289742201992, "darwin": 0.04389767248967019, "sagan": 0.12006669983039023, "hypatia": 0.20363755678568088, "socrates": 0.38660549352362394}, {"einstein": 0.22533815964505868, "sagan": 0.11150513598441134, "socrates": 0.22682554341020966, "ibn_sina": 0.08658036873292324, "darwin": 0.03670647907785871, "hypatia": 0.3130443131495383}, {"socrates": 0.3566695605750943, "ibn_sina": 0.10072553698926656, "hypatia": 0.29861945819192054, "einstein": 0.09594184517519985, "sagan": 0.10807032181535581, "darwin": 0.039973277253162964}], "prompt_tokens": 549, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.04599934472801319, "einstein": 0.11097956656593166, "hypatia": 0.2865316813630734, "ibn_sina": 0.09343128662340078, "sagan": 0.11288636275408764, "socrates": 0.3501717579654933}, "confidence": 0.08875096555349686}}, "usage": {"input_tokens": 549, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.15757634831267028, "hypatia": 0.3145041751457066, "sagan": 0.10632937716209864, "socrates": 0.2543279025451256, "darwin": 0.045411893242706496, "einstein": 0.12185030359169231}, "confidence": 0.059969844070713485}}, "usage": {"input_tokens": 549, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.23815634459996815, "darwin": 0.035158740506827706, "sagan": 0.12042827389905522, "ibn_sina": 0.07305273774954205, "einstein": 0.1623491375786141, "socrates": 0.3708547656659928}, "confidence": 0.09002170341274321}}, "usage": {"input_tokens": 549, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.14253967994861494, "einstein": 0.10325289742201992, "darwin": 0.04389767248967019, "sagan": 0.12006669983039023, "hypatia": 0.20363755678568088, "socrates": 0.38660549352362394}, "confidence": 0.08590456168042401}}, "usage": {"input_tokens": 549, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.22533815964505868, "sagan": 0.11150513598441134, "socrates": 0.22682554341020966, "ibn_sina": 0.08658036873292324, "darwin": 0.03670647907785871, "hypatia": 0.3130443131495383}, "confidence": 0.06580093184553922}}, "usage": {"input_tokens": 549, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.3566695605750943, "ibn_sina": 0.10072553698926656, "hypatia": 0.29861945819192054, "einstein": 0.09594184517519985, "sagan": 0.10807032181535581, "darwin": 0.039973277253162964}, "confidence": 0.09881714207283757}}, "usage": {"input_tokens": 549, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:05Z"}
{"id": "c-hypatia-01-dessert-02", "uid": "c038", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "prompt_sha256_by_order": ["cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70"], "choice": "sagan", "probs": {"darwin": 0.009874551220120634, "einstein": 0.006793815450503006, "hypatia": 0.00588720493887932, "ibn_sina": 0.006171357842845737, "sagan": 0.9547841276188713, "socrates": 0.016488942928780004}, "confidence": 0.8945212277459544, "seconds": 0.18892789899958493, "seconds_all_orders": 1.1335673939975095, "seconds_by_order": [0.1886735160005628, 0.18813939099891286, 0.18946611899991694, 0.1892605110006116, 0.18920746199910354, 0.18882039499840175], "argmax_stable": true, "spread": {"darwin": 0.005235352264857903, "einstein": 0.0021590516696967533, "hypatia": 0.0017420939900091182, "ibn_sina": 0.002441448445817708, "sagan": 0.010347340031847896, "socrates": 0.0063979517282117156}, "spread_2dp": {"darwin": 0.01, "einstein": 0.0, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.020000000000000018, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.009874551220120634, "einstein": 0.006793815450503006, "hypatia": 0.00588720493887932, "ibn_sina": 0.006171357842845737, "sagan": 0.9547841276188713, "socrates": 0.016488942928780004}, {"ibn_sina": 0.005555450713273572, "hypatia": 0.007629298928888438, "sagan": 0.9602998102592606, "socrates": 0.010090991200568288, "darwin": 0.008105111445784147, "einstein": 0.008319337452225}, {"hypatia": 0.006068532390579233, "darwin": 0.006879650390808811, "sagan": 0.9590837837529855, "ibn_sina": 0.005605755678255134, "einstein": 0.006399720578960625, "socrates": 0.015962557208410538}, {"ibn_sina": 0.005541079523116722, "einstein": 0.006160285782528247, "darwin": 0.008399439834181094, "sagan": 0.9571624111726582, "hypatia": 0.007481805234284415, "socrates": 0.01525497845323143}, {"einstein": 0.007482757221324761, "sagan": 0.9651314676507192, "socrates": 0.011501446157740314, "ibn_sina": 0.004705987621630646, "darwin": 0.004639198955262731, "hypatia": 0.006539142393322239}, {"socrates": 0.010767267684502878, "ibn_sina": 0.0037299093970280294, "hypatia": 0.006287931562881077, "einstein": 0.008065219867319865, "sagan": 0.9617279627337102, "darwin": 0.009421708754557993}], "prompt_tokens": 455, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.009874551220120634, "einstein": 0.006793815450503006, "hypatia": 0.00588720493887932, "ibn_sina": 0.006171357842845737, "sagan": 0.9547841276188713, "socrates": 0.016488942928780004}, "confidence": 0.8945212277459544}}, "usage": {"input_tokens": 455, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.005555450713273572, "hypatia": 0.007629298928888438, "sagan": 0.9602998102592606, "socrates": 0.010090991200568288, "darwin": 0.008105111445784147, "einstein": 0.008319337452225}, "confidence": 0.907001832555218}}, "usage": {"input_tokens": 455, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.006068532390579233, "darwin": 0.006879650390808811, "sagan": 0.9590837837529855, "ibn_sina": 0.005605755678255134, "einstein": 0.006399720578960625, "socrates": 0.015962557208410538}, "confidence": 0.9043036541027962}}, "usage": {"input_tokens": 455, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.005541079523116722, "einstein": 0.006160285782528247, "darwin": 0.008399439834181094, "sagan": 0.9571624111726582, "hypatia": 0.007481805234284415, "socrates": 0.01525497845323143}, "confidence": 0.8999053316941132}}, "usage": {"input_tokens": 455, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.007482757221324761, "sagan": 0.9651314676507192, "socrates": 0.011501446157740314, "ibn_sina": 0.004705987621630646, "darwin": 0.004639198955262731, "hypatia": 0.006539142393322239}, "confidence": 0.9181041443663986}}, "usage": {"input_tokens": 455, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.010767267684502878, "ibn_sina": 0.0037299093970280294, "hypatia": 0.006287931562881077, "einstein": 0.008065219867319865, "sagan": 0.9617279627337102, "darwin": 0.009421708754557993}, "confidence": 0.910292650038563}}, "usage": {"input_tokens": 455, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:06Z"}
{"id": "c-hypatia-01-dessert-03", "uid": "c039", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "prompt_sha256_by_order": ["a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e"], "choice": "socrates", "probs": {"darwin": 0.117053039249234, "einstein": 0.04185364646773353, "hypatia": 0.05972422508239382, "ibn_sina": 0.02130105216649834, "sagan": 0.13771893083242165, "socrates": 0.6223491062017187}, "confidence": 0.31091052820590936, "seconds": 0.18568851549995694, "seconds_all_orders": 1.1141310929997417, "seconds_by_order": [0.1851416949994018, 0.1858135200000106, 0.18648335400030192, 0.18591902799926174, 0.1856036210010643, 0.18516987499970128], "argmax_stable": false, "spread": {"darwin": 0.3585113881770552, "einstein": 0.05388377796071355, "hypatia": 0.21700486971073313, "ibn_sina": 0.011977386986035867, "sagan": 0.24860214129441424, "socrates": 0.5311549941905747}, "spread_2dp": {"darwin": 0.36, "einstein": 0.049999999999999996, "hypatia": 0.22000000000000003, "ibn_sina": 0.009999999999999998, "sagan": 0.25, "socrates": 0.53}, "choices_by_order": ["socrates", "darwin", "socrates", "socrates", "sagan", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.117053039249234, "einstein": 0.04185364646773353, "hypatia": 0.05972422508239382, "ibn_sina": 0.02130105216649834, "sagan": 0.13771893083242165, "socrates": 0.6223491062017187}, {"ibn_sina": 0.03327843915253421, "hypatia": 0.08429450507256128, "sagan": 0.09875164808882363, "socrates": 0.3187234748877568, "darwin": 0.4262064403431591, "einstein": 0.03874549245516493}, {"hypatia": 0.08746769816412986, "darwin": 0.08234550156295403, "sagan": 0.05164847836092271, "ibn_sina": 0.022339815122961217, "einstein": 0.038542595456271915, "socrates": 0.7176559113327603}, {"ibn_sina": 0.02386314460544205, "einstein": 0.06275561211836987, "darwin": 0.06769505216610391, "sagan": 0.05761485344415, "hypatia": 0.05750700584016426, "socrates": 0.7305643318257699}, {"einstein": 0.09242637341698547, "sagan": 0.30025061965533695, "socrates": 0.19940933763519517, "ibn_sina": 0.02619533519332884, "darwin": 0.11485965920711279, "hypatia": 0.26685867489204085}, {"socrates": 0.6050448775530386, "ibn_sina": 0.02735887056326649, "hypatia": 0.04985380518130772, "einstein": 0.04441410100648228, "sagan": 0.15119214691069505, "darwin": 0.12213619878520976}], "prompt_tokens": 503, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.117053039249234, "einstein": 0.04185364646773353, "hypatia": 0.05972422508239382, "ibn_sina": 0.02130105216649834, "sagan": 0.13771893083242165, "socrates": 0.6223491062017187}, "confidence": 0.31091052820590936}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.03327843915253421, "hypatia": 0.08429450507256128, "sagan": 0.09875164808882363, "socrates": 0.3187234748877568, "darwin": 0.4262064403431591, "einstein": 0.03874549245516493}, "confidence": 0.16324324302186694}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.08746769816412986, "darwin": 0.08234550156295403, "sagan": 0.05164847836092271, "ibn_sina": 0.022339815122961217, "einstein": 0.038542595456271915, "socrates": 0.7176559113327603}, "confidence": 0.4409362614902477}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.02386314460544205, "einstein": 0.06275561211836987, "darwin": 0.06769505216610391, "sagan": 0.05761485344415, "hypatia": 0.05750700584016426, "socrates": 0.7305643318257699}, "confidence": 0.4593293279256499}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.09242637341698547, "sagan": 0.30025061965533695, "socrates": 0.19940933763519517, "ibn_sina": 0.02619533519332884, "darwin": 0.11485965920711279, "hypatia": 0.26685867489204085}, "confidence": 0.06825957076808557}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.6050448775530386, "ibn_sina": 0.02735887056326649, "hypatia": 0.04985380518130772, "einstein": 0.04441410100648228, "sagan": 0.15119214691069505, "darwin": 0.12213619878520976}, "confidence": 0.29087457070252515}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:07Z"}
{"id": "c-hypatia-01-main-01", "uid": "c040", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "prompt_sha256_by_order": ["6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372"], "choice": "ibn_sina", "probs": {"darwin": 0.007719426015619184, "einstein": 0.019955807030724665, "hypatia": 0.33739936374458557, "ibn_sina": 0.6194624128740978, "sagan": 0.007420154509835068, "socrates": 0.008042835825137769}, "confidence": 0.3977794975508421, "seconds": 0.22666816833285944, "seconds_all_orders": 1.3600090099971567, "seconds_by_order": [0.2266597929992713, 0.22676739299822657, 0.22714338900004805, 0.22605139799998142, 0.22670152399950894, 0.2266855130001204], "argmax_stable": false, "spread": {"darwin": 0.002431423381796088, "einstein": 0.025393622448039443, "hypatia": 0.20598957866981382, "ibn_sina": 0.21493217039110657, "sagan": 0.003286777697335005, "socrates": 0.002830550842105109}, "spread_2dp": {"darwin": 0.0, "einstein": 0.03, "hypatia": 0.21000000000000002, "ibn_sina": 0.21000000000000002, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["ibn_sina", "hypatia", "ibn_sina", "hypatia", "hypatia", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.007719426015619184, "einstein": 0.019955807030724665, "hypatia": 0.33739936374458557, "ibn_sina": 0.6194624128740978, "sagan": 0.007420154509835068, "socrates": 0.008042835825137769}, {"ibn_sina": 0.40588238458925796, "hypatia": 0.5355605042612608, "sagan": 0.0078170134530886, "socrates": 0.005888210680017948, "darwin": 0.006693356081506173, "einstein": 0.038158530934868554}, {"hypatia": 0.32957092559144696, "darwin": 0.006480772163015897, "sagan": 0.009932706808329344, "ibn_sina": 0.6208145549803645, "einstein": 0.025812990154870475, "socrates": 0.007388050301972832}, {"ibn_sina": 0.4646907292860166, "einstein": 0.022777570917020252, "darwin": 0.006816865464831746, "sagan": 0.008832480335566011, "hypatia": 0.4881635924744423, "socrates": 0.008718761522123057}, {"einstein": 0.026476873141908126, "sagan": 0.009173149056789679, "socrates": 0.007654684489936574, "ibn_sina": 0.4696921603148483, "darwin": 0.008110662244722518, "hypatia": 0.4788924707517948}, {"socrates": 0.008285714365212884, "ibn_sina": 0.5622333899248052, "hypatia": 0.40439081924923204, "einstein": 0.012764908486829108, "sagan": 0.006645929110994339, "darwin": 0.00567923886292643}], "prompt_tokens": 641, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.007719426015619184, "einstein": 0.019955807030724665, "hypatia": 0.33739936374458557, "ibn_sina": 0.6194624128740978, "sagan": 0.007420154509835068, "socrates": 0.008042835825137769}, "confidence": 0.3977794975508421}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.40588238458925796, "hypatia": 0.5355605042612608, "sagan": 0.0078170134530886, "socrates": 0.005888210680017948, "darwin": 0.006693356081506173, "einstein": 0.038158530934868554}, "confidence": 0.343794658081021}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.32957092559144696, "darwin": 0.006480772163015897, "sagan": 0.009932706808329344, "ibn_sina": 0.6208145549803645, "einstein": 0.025812990154870475, "socrates": 0.007388050301972832}, "confidence": 0.39386711138924824}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.4646907292860166, "einstein": 0.022777570917020252, "darwin": 0.006816865464831746, "sagan": 0.008832480335566011, "hypatia": 0.4881635924744423, "socrates": 0.008718761522123057}, "confidence": 0.3459525805659121}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.026476873141908126, "sagan": 0.009173149056789679, "socrates": 0.007654684489936574, "ibn_sina": 0.4696921603148483, "darwin": 0.008110662244722518, "hypatia": 0.4788924707517948}, "confidence": 0.34102992701880036}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.008285714365212884, "ibn_sina": 0.5622333899248052, "hypatia": 0.40439081924923204, "einstein": 0.012764908486829108, "sagan": 0.006645929110994339, "darwin": 0.00567923886292643}, "confidence": 0.3759356050221192}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:09Z"}
{"id": "c-hypatia-01-main-01", "uid": "c041", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "prompt_sha256_by_order": ["c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c"], "choice": "einstein", "probs": {"darwin": 0.014511833015060243, "einstein": 0.6795680941668978, "hypatia": 0.05493484436785091, "ibn_sina": 0.010149639271177636, "sagan": 0.2278919623826794, "socrates": 0.01294362679633405}, "confidence": 0.4206958290440956, "seconds": 0.18527265783329008, "seconds_all_orders": 1.1116359469997406, "seconds_by_order": [0.18552558199917257, 0.18535122399953252, 0.18493516700073087, 0.18549815300139016, 0.18502040699968347, 0.18530541399923095], "argmax_stable": false, "spread": {"darwin": 0.014799212491513464, "einstein": 0.5367375550627125, "hypatia": 0.0277142411647274, "ibn_sina": 0.00469965745770125, "sagan": 0.5306227981793495, "socrates": 0.005393329907178777}, "spread_2dp": {"darwin": 0.019999999999999997, "einstein": 0.54, "hypatia": 0.030000000000000006, "ibn_sina": 0.0, "sagan": 0.53, "socrates": 0.0}, "choices_by_order": ["einstein", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.014511833015060243, "einstein": 0.6795680941668978, "hypatia": 0.05493484436785091, "ibn_sina": 0.010149639271177636, "sagan": 0.2278919623826794, "socrates": 0.01294362679633405}, {"ibn_sina": 0.010623701697966049, "hypatia": 0.07061191637760363, "sagan": 0.4713589051940489, "socrates": 0.010649847134117568, "darwin": 0.026147878378895463, "einstein": 0.41060775121736837}, {"hypatia": 0.05014660622293063, "darwin": 0.019462576215280983, "sagan": 0.5080690182421117, "ibn_sina": 0.01040656959225025, "einstein": 0.3986141078004034, "socrates": 0.01330112192702301}, {"ibn_sina": 0.0102321466800793, "einstein": 0.2270104422222827, "darwin": 0.029311045506573707, "sagan": 0.6795701837278442, "hypatia": 0.04289767521287623, "socrates": 0.010978506650343951}, {"einstein": 0.21261711603075703, "sagan": 0.6977165704216172, "socrates": 0.007907792019844232, "ibn_sina": 0.005924044240264799, "darwin": 0.02264094370825005, "hypatia": 0.05319353357926664}, {"socrates": 0.009758082313717376, "ibn_sina": 0.008844791580703006, "hypatia": 0.0567023293026476, "einstein": 0.14283053910418525, "sagan": 0.7585147605620289, "darwin": 0.023349497136717846}], "prompt_tokens": 497, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.014511833015060243, "einstein": 0.6795680941668978, "hypatia": 0.05493484436785091, "ibn_sina": 0.010149639271177636, "sagan": 0.2278919623826794, "socrates": 0.01294362679633405}, "confidence": 0.4206958290440956}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.010623701697966049, "hypatia": 0.07061191637760363, "sagan": 0.4713589051940489, "socrates": 0.010649847134117568, "darwin": 0.026147878378895463, "einstein": 0.41060775121736837}, "confidence": 0.27600877531059914}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.05014660622293063, "darwin": 0.019462576215280983, "sagan": 0.5080690182421117, "ibn_sina": 0.01040656959225025, "einstein": 0.3986141078004034, "socrates": 0.01330112192702301}, "confidence": 0.3042472297105888}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.0102321466800793, "einstein": 0.2270104422222827, "darwin": 0.029311045506573707, "sagan": 0.6795701837278442, "hypatia": 0.04289767521287623, "socrates": 0.010978506650343951}, "confidence": 0.4195287374213973}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.21261711603075703, "sagan": 0.6977165704216172, "socrates": 0.007907792019844232, "ibn_sina": 0.005924044240264799, "darwin": 0.02264094370825005, "hypatia": 0.05319353357926664}, "confidence": 0.4425450909898939}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.009758082313717376, "ibn_sina": 0.008844791580703006, "hypatia": 0.0567023293026476, "einstein": 0.14283053910418525, "sagan": 0.7585147605620289, "darwin": 0.023349497136717846}, "confidence": 0.5196148102776341}}, "usage": {"input_tokens": 497, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:10Z"}
{"id": "c-hypatia-01-main-01", "uid": "c042", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "prompt_sha256_by_order": ["79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60"], "choice": "hypatia", "probs": {"darwin": 0.023179254898770513, "einstein": 0.19191319996524456, "hypatia": 0.44651862634808825, "ibn_sina": 0.13318114597180902, "sagan": 0.03195357654887601, "socrates": 0.17325419626721167}, "confidence": 0.14262652369020695, "seconds": 0.2308388104996387, "seconds_all_orders": 1.3850328629978321, "seconds_by_order": [0.23062804999972286, 0.23086048900040623, 0.23056652099876374, 0.23122479599987855, 0.23096359800001665, 0.2307894089990441], "argmax_stable": true, "spread": {"darwin": 0.009143753526447931, "einstein": 0.11230755340148374, "hypatia": 0.13458292734002653, "ibn_sina": 0.057522992664958195, "sagan": 0.025817181811152924, "socrates": 0.08963800950817674}, "spread_2dp": {"darwin": 0.009999999999999998, "einstein": 0.10999999999999999, "hypatia": 0.12999999999999995, "ibn_sina": 0.05, "sagan": 0.020000000000000004, "socrates": 0.09000000000000001}, "choices_by_order": ["hypatia", "hypatia", "hypatia", "hypatia", "hypatia", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.023179254898770513, "einstein": 0.19191319996524456, "hypatia": 0.44651862634808825, "ibn_sina": 0.13318114597180902, "sagan": 0.03195357654887601, "socrates": 0.17325419626721167}, {"ibn_sina": 0.10507034383531472, "hypatia": 0.5811015536881148, "sagan": 0.026751178172518302, "socrates": 0.08361618675903493, "darwin": 0.02880656873602562, "einstein": 0.1746541688089916}, {"hypatia": 0.4760529328527922, "darwin": 0.01966281520957769, "sagan": 0.026016683630412887, "ibn_sina": 0.1625933365002729, "einstein": 0.197034526027322, "socrates": 0.11863970577962235}, {"ibn_sina": 0.13269425517145197, "einstein": 0.24466336913297307, "darwin": 0.02143700915706924, "sagan": 0.03573803996115982, "hypatia": 0.46897162922408986, "socrates": 0.096495697353256}, {"einstein": 0.19779491383570472, "sagan": 0.05183386544156581, "socrates": 0.11277051759357928, "ibn_sina": 0.13999986912287934, "darwin": 0.023771805744225855, "hypatia": 0.473829028262045}, {"socrates": 0.1624142378400566, "ibn_sina": 0.16226021788639305, "hypatia": 0.47696637407937853, "einstein": 0.13235581573148933, "sagan": 0.04137780339782748, "darwin": 0.02462555106485504}], "prompt_tokens": 677, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"darwin": 0.023179254898770513, "einstein": 0.19191319996524456, "hypatia": 0.44651862634808825, "ibn_sina": 0.13318114597180902, "sagan": 0.03195357654887601, "socrates": 0.17325419626721167}, "confidence": 0.14262652369020695}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.10507034383531472, "hypatia": 0.5811015536881148, "sagan": 0.026751178172518302, "socrates": 0.08361618675903493, "darwin": 0.02880656873602562, "einstein": 0.1746541688089916}, "confidence": 0.26531197859062594}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.4760529328527922, "darwin": 0.01966281520957769, "sagan": 0.026016683630412887, "ibn_sina": 0.1625933365002729, "einstein": 0.197034526027322, "socrates": 0.11863970577962235}, "confidence": 0.16842935957849664}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.13269425517145197, "einstein": 0.24466336913297307, "darwin": 0.02143700915706924, "sagan": 0.03573803996115982, "hypatia": 0.46897162922408986, "socrates": 0.096495697353256}, "confidence": 0.17014058924518957}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.19779491383570472, "sagan": 0.05183386544156581, "socrates": 0.11277051759357928, "ibn_sina": 0.13999986912287934, "darwin": 0.023771805744225855, "hypatia": 0.473829028262045}, "confidence": 0.15904693277287638}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"socrates": 0.1624142378400566, "ibn_sina": 0.16226021788639305, "hypatia": 0.47696637407937853, "einstein": 0.13235581573148933, "sagan": 0.04137780339782748, "darwin": 0.02462555106485504}, "confidence": 0.16004834476163307}}, "usage": {"input_tokens": 677, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:11Z"}
{"id": "c-hypatia-01-main-01", "uid": "c043", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "prompt_sha256_by_order": ["1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568"], "choice": "sagan", "probs": {"darwin": 0.032045744963649626, "einstein": 0.21995435389347762, "hypatia": 0.08405933142382949, "ibn_sina": 0.18738549163756196, "sagan": 0.456793564096752, "socrates": 0.019761513984729324}, "confidence": 0.16076442265339191, "seconds": 0.22120810549949965, "seconds_all_orders": 1.327248632996998, "seconds_by_order": [0.22082591199978197, 0.2215791859998717, 0.22094803099935234, 0.2212033390005672, 0.22149586599880422, 0.22119629899862048], "argmax_stable": true, "spread": {"darwin": 0.019289293845601532, "einstein": 0.060782681956104756, "hypatia": 0.03370022069080118, "ibn_sina": 0.10073828031592177, "sagan": 0.10592825459107647, "socrates": 0.0021359789145835827}, "spread_2dp": {"darwin": 0.020000000000000004, "einstein": 0.06, "hypatia": 0.04000000000000001, "ibn_sina": 0.1, "sagan": 0.10000000000000003, "socrates": 0.0}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.032045744963649626, "einstein": 0.21995435389347762, "hypatia": 0.08405933142382949, "ibn_sina": 0.18738549163756196, "sagan": 0.456793564096752, "socrates": 0.019761513984729324}, {"ibn_sina": 0.09755032249348355, "hypatia": 0.09747370797918067, "sagan": 0.46976884198394775, "socrates": 0.021875499274806337, "darwin": 0.05133503880925116, "einstein": 0.26199658945933063}, {"hypatia": 0.06377348728837949, "darwin": 0.032152713017847515, "sagan": 0.5627218186878284, "ibn_sina": 0.11713389084396394, "einstein": 0.2023205972626677, "socrates": 0.021897492899312906}, {"ibn_sina": 0.0877881614131978, "einstein": 0.26310327921877247, "darwin": 0.03898505828689708, "sagan": 0.5216682912886449, "hypatia": 0.06710631909665435, "socrates": 0.021348890695833365}, {"einstein": 0.22688008170326757, "sagan": 0.531634573409656, "socrates": 0.021093448268614493, "ibn_sina": 0.1090747054714189, "darwin": 0.03722351797193188, "hypatia": 0.07409367317511115}, {"socrates": 0.021881833489926783, "ibn_sina": 0.08664721132164019, "hypatia": 0.0926870614403321, "einstein": 0.21909748764318263, "sagan": 0.5409629139783139, "darwin": 0.03872349212660442}], "prompt_tokens": 581, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.032045744963649626, "einstein": 0.21995435389347762, "hypatia": 0.08405933142382949, "ibn_sina": 0.18738549163756196, "sagan": 0.456793564096752, "socrates": 0.019761513984729324}, "confidence": 0.16076442265339191}}, "usage": {"input_tokens": 581, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.09755032249348355, "hypatia": 0.09747370797918067, "sagan": 0.46976884198394775, "socrates": 0.021875499274806337, "darwin": 0.05133503880925116, "einstein": 0.26199658945933063}, "confidence": 0.17374718875734746}}, "usage": {"input_tokens": 581, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.06377348728837949, "darwin": 0.032152713017847515, "sagan": 0.5627218186878284, "ibn_sina": 0.11713389084396394, "einstein": 0.2023205972626677, "socrates": 0.021897492899312906}, "confidence": 0.2522682070226921}}, "usage": {"input_tokens": 581, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.0877881614131978, "einstein": 0.26310327921877247, "darwin": 0.03898505828689708, "sagan": 0.5216682912886449, "hypatia": 0.06710631909665435, "socrates": 0.021348890695833365}, "confidence": 0.22665612510672747}}, "usage": {"input_tokens": 581, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.22688008170326757, "sagan": 0.531634573409656, "socrates": 0.021093448268614493, "ibn_sina": 0.1090747054714189, "darwin": 0.03722351797193188, "hypatia": 0.07409367317511115}, "confidence": 0.22399309449576724}}, "usage": {"input_tokens": 581, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.021881833489926783, "ibn_sina": 0.08664721132164019, "hypatia": 0.0926870614403321, "einstein": 0.21909748764318263, "sagan": 0.5409629139783139, "darwin": 0.03872349212660442}, "confidence": 0.23046584495090094}}, "usage": {"input_tokens": 581, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:13Z"}
{"id": "c-hypatia-01-main-01", "uid": "c044", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "prompt_sha256_by_order": ["550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0"], "choice": "sagan", "probs": {"darwin": 0.08363705876892295, "einstein": 0.06277240290910016, "hypatia": 0.06942358015355003, "ibn_sina": 0.030693248424190137, "sagan": 0.6288991794269281, "socrates": 0.12457453031730861}, "confidence": 0.31327627916116685, "seconds": 0.18950720133367818, "seconds_all_orders": 1.1370432080020692, "seconds_by_order": [0.18964190900078393, 0.18910746300025494, 0.1898336369995377, 0.1898300670000026, 0.18904442300117807, 0.1895857090003119], "argmax_stable": true, "spread": {"darwin": 0.04488057239525836, "einstein": 0.021745226367624153, "hypatia": 0.03493820057315464, "ibn_sina": 0.008691558918505694, "sagan": 0.034188383099167274, "socrates": 0.02701312230559752}, "spread_2dp": {"darwin": 0.039999999999999994, "einstein": 0.019999999999999997, "hypatia": 0.039999999999999994, "ibn_sina": 0.009999999999999998, "sagan": 0.030000000000000027, "socrates": 0.01999999999999999}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.08363705876892295, "einstein": 0.06277240290910016, "hypatia": 0.06942358015355003, "ibn_sina": 0.030693248424190137, "sagan": 0.6288991794269281, "socrates": 0.12457453031730861}, {"ibn_sina": 0.027044022327818475, "hypatia": 0.07696253063594788, "sagan": 0.636998652881381, "socrates": 0.10394795119282974, "darwin": 0.11329218287958136, "einstein": 0.041754660082441504}, {"hypatia": 0.05258689649628384, "darwin": 0.068411610484323, "sagan": 0.6630875625260954, "ibn_sina": 0.028358845196589717, "einstein": 0.06349988645006566, "socrates": 0.12405519884664247}, {"ibn_sina": 0.02941710969174044, "einstein": 0.055606752863037394, "darwin": 0.08768618578562067, "sagan": 0.6334963545494021, "hypatia": 0.06945995341648067, "socrates": 0.12433364369371869}, {"einstein": 0.05883396023687225, "sagan": 0.6529258951804143, "socrates": 0.09756140801171109, "ibn_sina": 0.030696741725724095, "darwin": 0.07245689777583975, "hypatia": 0.08752509706943848}, {"socrates": 0.1162860052323833, "ibn_sina": 0.0220051828072184, "hypatia": 0.0610148835625176, "einstein": 0.061457535694315654, "sagan": 0.6378258951190559, "darwin": 0.10141049758450915}], "prompt_tokens": 457, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.08363705876892295, "einstein": 0.06277240290910016, "hypatia": 0.06942358015355003, "ibn_sina": 0.030693248424190137, "sagan": 0.6288991794269281, "socrates": 0.12457453031730861}, "confidence": 0.31327627916116685}}, "usage": {"input_tokens": 457, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.027044022327818475, "hypatia": 0.07696253063594788, "sagan": 0.636998652881381, "socrates": 0.10394795119282974, "darwin": 0.11329218287958136, "einstein": 0.041754660082441504}, "confidence": 0.32536676912279033}}, "usage": {"input_tokens": 457, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.05258689649628384, "darwin": 0.068411610484323, "sagan": 0.6630875625260954, "ibn_sina": 0.028358845196589717, "einstein": 0.06349988645006566, "socrates": 0.12405519884664247}, "confidence": 0.36082815729988404}}, "usage": {"input_tokens": 457, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.02941710969174044, "einstein": 0.055606752863037394, "darwin": 0.08768618578562067, "sagan": 0.6334963545494021, "hypatia": 0.06945995341648067, "socrates": 0.12433364369371869}, "confidence": 0.3198970189532503}}, "usage": {"input_tokens": 457, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.05883396023687225, "sagan": 0.6529258951804143, "socrates": 0.09756140801171109, "ibn_sina": 0.030696741725724095, "darwin": 0.07245689777583975, "hypatia": 0.08752509706943848}, "confidence": 0.34377378689489896}}, "usage": {"input_tokens": 457, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.1162860052323833, "ibn_sina": 0.0220051828072184, "hypatia": 0.0610148835625176, "einstein": 0.061457535694315654, "sagan": 0.6378258951190559, "darwin": 0.10141049758450915}, "confidence": 0.32633496315737587}}, "usage": {"input_tokens": 457, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:14Z"}
{"id": "c-hypatia-01-main-02", "uid": "c045", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "prompt_sha256_by_order": ["a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2"], "choice": "ibn_sina", "probs": {"darwin": 0.007528263213800016, "einstein": 0.13047928376447182, "hypatia": 0.1580784562863926, "ibn_sina": 0.6757310073777802, "sagan": 0.01742215289422897, "socrates": 0.010760836463326392}, "confidence": 0.39892244551029366, "seconds": 0.22347618816608397, "seconds_all_orders": 1.3408571289965039, "seconds_by_order": [0.22361796899895126, 0.22404657500010217, 0.22286531499958073, 0.2232649509987823, 0.22350914999879024, 0.2235531690002972], "argmax_stable": true, "spread": {"darwin": 0.002365106439083501, "einstein": 0.198718641054571, "hypatia": 0.18087817210212978, "ibn_sina": 0.33026835714190794, "sagan": 0.009900873850712883, "socrates": 0.004392258886154467}, "spread_2dp": {"darwin": 0.0, "einstein": 0.2, "hypatia": 0.19, "ibn_sina": 0.33000000000000007, "sagan": 0.01, "socrates": 0.0}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.007528263213800016, "einstein": 0.13047928376447182, "hypatia": 0.1580784562863926, "ibn_sina": 0.6757310073777802, "sagan": 0.01742215289422897, "socrates": 0.010760836463326392}, {"ibn_sina": 0.6169237951405031, "hypatia": 0.11443018692272741, "sagan": 0.007521279043516086, "socrates": 0.009777057227521804, "darwin": 0.0074123118538346865, "einstein": 0.24393536981189698}, {"hypatia": 0.28991930965572404, "darwin": 0.006128634072569434, "sagan": 0.008103817962116242, "ibn_sina": 0.43267663056363614, "einstein": 0.25111707644746856, "socrates": 0.012054531298485589}, {"ibn_sina": 0.5469147919140551, "einstein": 0.14173684578633608, "darwin": 0.006065082966469239, "sagan": 0.007857576506940382, "hypatia": 0.28327364707758984, "socrates": 0.014152055748609272}, {"einstein": 0.3291979248190428, "sagan": 0.014202084674166483, "socrates": 0.009759796862454805, "ibn_sina": 0.3454626502358723, "darwin": 0.006069184383606454, "hypatia": 0.2953083590248572}, {"socrates": 0.010386356833402586, "ibn_sina": 0.6660985564838525, "hypatia": 0.14459946457319844, "einstein": 0.16403843938773752, "sagan": 0.009714025947092433, "darwin": 0.005163156774716515}], "prompt_tokens": 605, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.007528263213800016, "einstein": 0.13047928376447182, "hypatia": 0.1580784562863926, "ibn_sina": 0.6757310073777802, "sagan": 0.01742215289422897, "socrates": 0.010760836463326392}, "confidence": 0.39892244551029366}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.6169237951405031, "hypatia": 0.11443018692272741, "sagan": 0.007521279043516086, "socrates": 0.009777057227521804, "darwin": 0.0074123118538346865, "einstein": 0.24393536981189698}, "confidence": 0.34408096502619745}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.28991930965572404, "darwin": 0.006128634072569434, "sagan": 0.008103817962116242, "ibn_sina": 0.43267663056363614, "einstein": 0.25111707644746856, "socrates": 0.012054531298485589}, "confidence": 0.2014847230919937}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.5469147919140551, "einstein": 0.14173684578633608, "darwin": 0.006065082966469239, "sagan": 0.007857576506940382, "hypatia": 0.28327364707758984, "socrates": 0.014152055748609272}, "confidence": 0.27969746754202846}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.3291979248190428, "sagan": 0.014202084674166483, "socrates": 0.009759796862454805, "ibn_sina": 0.3454626502358723, "darwin": 0.006069184383606454, "hypatia": 0.2953083590248572}, "confidence": 0.17830783739929784}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.010386356833402586, "ibn_sina": 0.6660985564838525, "hypatia": 0.14459946457319844, "einstein": 0.16403843938773752, "sagan": 0.009714025947092433, "darwin": 0.005163156774716515}, "confidence": 0.3900805583172929}}, "usage": {"input_tokens": 605, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:15Z"}
{"id": "c-hypatia-01-main-03", "uid": "c046", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "prompt_sha256_by_order": ["51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe"], "choice": "ibn_sina", "probs": {"darwin": 0.009934163856987506, "einstein": 0.2426904450701945, "hypatia": 0.23062361382635269, "ibn_sina": 0.3801701714771766, "sagan": 0.008526733315083045, "socrates": 0.12805487245420571}, "confidence": 0.1278216469786102, "seconds": 0.22591664116680477, "seconds_all_orders": 1.3554998470008286, "seconds_by_order": [0.22631165600068925, 0.22518781600047078, 0.2261275580003712, 0.2260258280002745, 0.22638177599947085, 0.22546521299955202], "argmax_stable": false, "spread": {"darwin": 0.003256174287703021, "einstein": 0.34276511480808947, "hypatia": 0.29000618069561446, "ibn_sina": 0.457459107274321, "sagan": 0.004293897719661972, "socrates": 0.16143304879570453}, "spread_2dp": {"darwin": 0.0, "einstein": 0.34, "hypatia": 0.29000000000000004, "ibn_sina": 0.45, "sagan": 0.0, "socrates": 0.16}, "choices_by_order": ["ibn_sina", "hypatia", "hypatia", "hypatia", "hypatia", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.009934163856987506, "einstein": 0.2426904450701945, "hypatia": 0.23062361382635269, "ibn_sina": 0.3801701714771766, "sagan": 0.008526733315083045, "socrates": 0.12805487245420571}, {"ibn_sina": 0.21414031071418224, "hypatia": 0.37396858862281146, "sagan": 0.008340627490814, "socrates": 0.03288064802082216, "darwin": 0.010114314321755243, "einstein": 0.3605555108296149}, {"hypatia": 0.5206297945219671, "darwin": 0.009310879770990196, "sagan": 0.009864701869715872, "ibn_sina": 0.11240960798692294, "einstein": 0.278976419051563, "socrates": 0.0688085967988408}, {"ibn_sina": 0.16855503435294036, "einstein": 0.13847825564796237, "darwin": 0.006858140034052222, "sagan": 0.007960138161796011, "hypatia": 0.48383473498672236, "socrates": 0.1943136968165267}, {"einstein": 0.3924514227216183, "sagan": 0.012254035881457983, "socrates": 0.0443480357814588, "ibn_sina": 0.076499513802145, "darwin": 0.009419089009749596, "hypatia": 0.4650279028035703}, {"socrates": 0.1381076462672833, "ibn_sina": 0.533958621076466, "hypatia": 0.25973416897575247, "einstein": 0.04968630791352884, "sagan": 0.010719374117479117, "darwin": 0.0077938816494902904}], "prompt_tokens": 631, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.009934163856987506, "einstein": 0.2426904450701945, "hypatia": 0.23062361382635269, "ibn_sina": 0.3801701714771766, "sagan": 0.008526733315083045, "socrates": 0.12805487245420571}, "confidence": 0.1278216469786102}}, "usage": {"input_tokens": 631, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.21414031071418224, "hypatia": 0.37396858862281146, "sagan": 0.008340627490814, "socrates": 0.03288064802082216, "darwin": 0.010114314321755243, "einstein": 0.3605555108296149}, "confidence": 0.18035422812918114}}, "usage": {"input_tokens": 631, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.5206297945219671, "darwin": 0.009310879770990196, "sagan": 0.009864701869715872, "ibn_sina": 0.11240960798692294, "einstein": 0.278976419051563, "socrates": 0.0688085967988408}, "confidence": 0.23972532774052757}}, "usage": {"input_tokens": 631, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.16855503435294036, "einstein": 0.13847825564796237, "darwin": 0.006858140034052222, "sagan": 0.007960138161796011, "hypatia": 0.48383473498672236, "socrates": 0.1943136968165267}, "confidence": 0.18346154599297818}}, "usage": {"input_tokens": 631, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.3924514227216183, "sagan": 0.012254035881457983, "socrates": 0.0443480357814588, "ibn_sina": 0.076499513802145, "darwin": 0.009419089009749596, "hypatia": 0.4650279028035703}, "confidence": 0.253992248925886}}, "usage": {"input_tokens": 631, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.1381076462672833, "ibn_sina": 0.533958621076466, "hypatia": 0.25973416897575247, "einstein": 0.04968630791352884, "sagan": 0.010719374117479117, "darwin": 0.0077938816494902904}, "confidence": 0.24915009793553144}}, "usage": {"input_tokens": 631, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:16Z"}
{"id": "c-hypatia-01-main-03", "uid": "c047", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "prompt_sha256_by_order": ["ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f"], "choice": "einstein", "probs": {"darwin": 0.011084421561859084, "einstein": 0.6429122831769702, "hypatia": 0.0487096267083883, "ibn_sina": 0.015607429069442558, "sagan": 0.27357730060241603, "socrates": 0.008108938880923714}, "confidence": 0.38918269855851195, "seconds": 0.21960754483370692, "seconds_all_orders": 1.3176452690022415, "seconds_by_order": [0.21972976100005326, 0.21931102400048985, 0.22009961700132408, 0.218829977000496, 0.21981049999885727, 0.2198643900010211], "argmax_stable": true, "spread": {"darwin": 0.00628804569212475, "einstein": 0.07690221529227581, "hypatia": 0.022087277894285023, "ibn_sina": 0.006648443944672039, "sagan": 0.09169434378009239, "socrates": 0.0033090644439583514}, "spread_2dp": {"darwin": 0.0, "einstein": 0.07000000000000006, "hypatia": 0.020000000000000004, "ibn_sina": 0.01, "sagan": 0.09999999999999998, "socrates": 0.0}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.011084421561859084, "einstein": 0.6429122831769702, "hypatia": 0.0487096267083883, "ibn_sina": 0.015607429069442558, "sagan": 0.27357730060241603, "socrates": 0.008108938880923714}, {"ibn_sina": 0.010131286492873804, "hypatia": 0.026651347704471105, "sagan": 0.35411341805506685, "socrates": 0.010341440930574052, "darwin": 0.01120768630296257, "einstein": 0.5875548205140515}, {"hypatia": 0.03704808807329116, "darwin": 0.008081845098057244, "sagan": 0.3155962185275737, "ibn_sina": 0.013232075683878382, "einstein": 0.6163966593846748, "socrates": 0.009645113232524772}, {"ibn_sina": 0.01106973642071549, "einstein": 0.6380604635837338, "darwin": 0.01282765806789938, "sagan": 0.30052270387579033, "hypatia": 0.02977662435671917, "socrates": 0.007742813695141862}, {"einstein": 0.5828784322892541, "sagan": 0.3443953539868332, "socrates": 0.008488990624611493, "ibn_sina": 0.00895898512477052, "darwin": 0.0065396123757746295, "hypatia": 0.04873862559875613}, {"socrates": 0.007032376486615701, "ibn_sina": 0.012858980407464166, "hypatia": 0.03858299414840692, "einstein": 0.5660100678846944, "sagan": 0.3652716443825084, "darwin": 0.010243936690310353}], "prompt_tokens": 565, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.011084421561859084, "einstein": 0.6429122831769702, "hypatia": 0.0487096267083883, "ibn_sina": 0.015607429069442558, "sagan": 0.27357730060241603, "socrates": 0.008108938880923714}, "confidence": 0.38918269855851195}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.010131286492873804, "hypatia": 0.026651347704471105, "sagan": 0.35411341805506685, "socrates": 0.010341440930574052, "darwin": 0.01120768630296257, "einstein": 0.5875548205140515}, "confidence": 0.3659949698670319}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.03704808807329116, "darwin": 0.008081845098057244, "sagan": 0.3155962185275737, "ibn_sina": 0.013232075683878382, "einstein": 0.6163966593846748, "socrates": 0.009645113232524772}, "confidence": 0.377502169522886}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.01106973642071549, "einstein": 0.6380604635837338, "darwin": 0.01282765806789938, "sagan": 0.30052270387579033, "hypatia": 0.02977662435671917, "socrates": 0.007742813695141862}, "confidence": 0.3984024845577222}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.5828784322892541, "sagan": 0.3443953539868332, "socrates": 0.008488990624611493, "ibn_sina": 0.00895898512477052, "darwin": 0.0065396123757746295, "hypatia": 0.04873862559875613}, "confidence": 0.35311116784871965}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.007032376486615701, "ibn_sina": 0.012858980407464166, "hypatia": 0.03858299414840692, "einstein": 0.5660100678846944, "sagan": 0.3652716443825084, "darwin": 0.010243936690310353}, "confidence": 0.3467189974111339}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:18Z"}
{"id": "c-hypatia-01-main-03", "uid": "c048", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "prompt_sha256_by_order": ["518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec"], "choice": "hypatia", "probs": {"darwin": 0.007224287606494035, "einstein": 0.058521579645145774, "hypatia": 0.8117906963658401, "ibn_sina": 0.053274049692823205, "sagan": 0.06296650496909102, "socrates": 0.0062228817206059165}, "confidence": 0.6031872756367896, "seconds": 0.2562063353328388, "seconds_all_orders": 1.537238011997033, "seconds_by_order": [0.25655831699987175, 0.2560001519996149, 0.2557293739992019, 0.2555305360001512, 0.2570178139994823, 0.25640181899871095], "argmax_stable": true, "spread": {"darwin": 0.0017268102992308524, "einstein": 0.1110721466614669, "hypatia": 0.1192006617842285, "ibn_sina": 0.023788444409260884, "sagan": 0.03987116127721041, "socrates": 0.0025488995136461216}, "spread_2dp": {"darwin": 0.0, "einstein": 0.11000000000000001, "hypatia": 0.10999999999999999, "ibn_sina": 0.020000000000000004, "sagan": 0.039999999999999994, "socrates": 0.0}, "choices_by_order": ["hypatia", "hypatia", "hypatia", "hypatia", "hypatia", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.007224287606494035, "einstein": 0.058521579645145774, "hypatia": 0.8117906963658401, "ibn_sina": 0.053274049692823205, "sagan": 0.06296650496909102, "socrates": 0.0062228817206059165}, {"ibn_sina": 0.02948560528356232, "hypatia": 0.8061446755327579, "sagan": 0.025780839762619034, "socrates": 0.0076526230025163984, "darwin": 0.005797559471967362, "einstein": 0.12513869694657692}, {"hypatia": 0.7890999659764797, "darwin": 0.006440537274348965, "sagan": 0.02448824642550873, "ibn_sina": 0.045959264920817686, "einstein": 0.12890826191397467, "socrates": 0.005103723488870277}, {"ibn_sina": 0.04666170344554628, "einstein": 0.0661637440666294, "darwin": 0.007524369771198214, "sagan": 0.03854224856054058, "hypatia": 0.8342447218249351, "socrates": 0.006863212331150354}, {"einstein": 0.16959372630661268, "sagan": 0.06435940770271914, "socrates": 0.006573271822931923, "ibn_sina": 0.03738565938064213, "darwin": 0.007043874746387464, "hypatia": 0.7150440600407066}, {"socrates": 0.005766472398185633, "ibn_sina": 0.0449970336836148, "hypatia": 0.8071058410479002, "einstein": 0.10654568572421864, "sagan": 0.028544038936366506, "darwin": 0.007040928209714281}], "prompt_tokens": 719, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"darwin": 0.007224287606494035, "einstein": 0.058521579645145774, "hypatia": 0.8117906963658401, "ibn_sina": 0.053274049692823205, "sagan": 0.06296650496909102, "socrates": 0.0062228817206059165}, "confidence": 0.6031872756367896}}, "usage": {"input_tokens": 719, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.02948560528356232, "hypatia": 0.8061446755327579, "sagan": 0.025780839762619034, "socrates": 0.0076526230025163984, "darwin": 0.005797559471967362, "einstein": 0.12513869694657692}, "confidence": 0.6005861899789077}}, "usage": {"input_tokens": 719, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.7890999659764797, "darwin": 0.006440537274348965, "sagan": 0.02448824642550873, "ibn_sina": 0.045959264920817686, "einstein": 0.12890826191397467, "socrates": 0.005103723488870277}, "confidence": 0.5704906636631857}}, "usage": {"input_tokens": 719, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.04666170344554628, "einstein": 0.0661637440666294, "darwin": 0.007524369771198214, "sagan": 0.03854224856054058, "hypatia": 0.8342447218249351, "socrates": 0.006863212331150354}, "confidence": 0.6449301234858119}}, "usage": {"input_tokens": 719, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.16959372630661268, "sagan": 0.06435940770271914, "socrates": 0.006573271822931923, "ibn_sina": 0.03738565938064213, "darwin": 0.007043874746387464, "hypatia": 0.7150440600407066}, "confidence": 0.4548192217157573}}, "usage": {"input_tokens": 719, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"socrates": 0.005766472398185633, "ibn_sina": 0.0449970336836148, "hypatia": 0.8071058410479002, "einstein": 0.10654568572421864, "sagan": 0.028544038936366506, "darwin": 0.007040928209714281}, "confidence": 0.598832972647817}}, "usage": {"input_tokens": 719, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:19Z"}
{"id": "c-hypatia-01-starter-01", "uid": "c049", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "prompt_sha256_by_order": ["2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e"], "choice": "sagan", "probs": {"darwin": 0.037532030976175794, "einstein": 0.026580600904163997, "hypatia": 0.03299158910876974, "ibn_sina": 0.050720871889325045, "sagan": 0.48318284216370144, "socrates": 0.3689920649578639}, "confidence": 0.25047644374482303, "seconds": 0.18888098416634116, "seconds_all_orders": 1.133285904998047, "seconds_by_order": [0.18854242700035684, 0.1884155779989669, 0.18854855700010376, 0.18986495600074704, 0.18899263299863378, 0.18892175399923872], "argmax_stable": false, "spread": {"darwin": 0.023595369617064852, "einstein": 0.009428340877343216, "hypatia": 0.01288593260137745, "ibn_sina": 0.03351838722447139, "sagan": 0.5495760565198562, "socrates": 0.5741912994316647}, "spread_2dp": {"darwin": 0.03, "einstein": 0.009999999999999998, "hypatia": 0.009999999999999998, "ibn_sina": 0.030000000000000002, "sagan": 0.5499999999999999, "socrates": 0.5700000000000001}, "choices_by_order": ["sagan", "sagan", "socrates", "socrates", "sagan", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.037532030976175794, "einstein": 0.026580600904163997, "hypatia": 0.03299158910876974, "ibn_sina": 0.050720871889325045, "sagan": 0.48318284216370144, "socrates": 0.3689920649578639}, {"ibn_sina": 0.033766185639808884, "hypatia": 0.026519489353822367, "sagan": 0.6324228963850327, "socrates": 0.2718149504973587, "darwin": 0.013936661359110942, "einstein": 0.021539816764866396}, {"hypatia": 0.020105656507392287, "darwin": 0.015258453780850667, "sagan": 0.39102188421654066, "ibn_sina": 0.02871139007349792, "einstein": 0.020747292932358316, "socrates": 0.5241553224893601}, {"ibn_sina": 0.017202484664853655, "einstein": 0.01715226002682078, "darwin": 0.017058578257824544, "sagan": 0.14435505318596387, "hypatia": 0.03263929040872489, "socrates": 0.7715923334558124}, {"einstein": 0.024149414883986524, "sagan": 0.6939311097058201, "socrates": 0.19740103402414771, "ibn_sina": 0.03407507211902187, "darwin": 0.020004439941672565, "hypatia": 0.03043892932535118}, {"socrates": 0.5050608983752651, "ibn_sina": 0.020522407004047442, "hypatia": 0.031968449401848544, "einstein": 0.025089140137675197, "sagan": 0.39163374598956163, "darwin": 0.025725359091602014}], "prompt_tokens": 451, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.037532030976175794, "einstein": 0.026580600904163997, "hypatia": 0.03299158910876974, "ibn_sina": 0.050720871889325045, "sagan": 0.48318284216370144, "socrates": 0.3689920649578639}, "confidence": 0.25047644374482303}}, "usage": {"input_tokens": 451, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.033766185639808884, "hypatia": 0.026519489353822367, "sagan": 0.6324228963850327, "socrates": 0.2718149504973587, "darwin": 0.013936661359110942, "einstein": 0.021539816764866396}, "confidence": 0.3716124640363209}}, "usage": {"input_tokens": 451, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.020105656507392287, "darwin": 0.015258453780850667, "sagan": 0.39102188421654066, "ibn_sina": 0.02871139007349792, "einstein": 0.020747292932358316, "socrates": 0.5241553224893601}, "confidence": 0.3154345215394338}}, "usage": {"input_tokens": 451, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.017202484664853655, "einstein": 0.01715226002682078, "darwin": 0.017058578257824544, "sagan": 0.14435505318596387, "hypatia": 0.03263929040872489, "socrates": 0.7715923334558124}, "confidence": 0.5417674651616122}}, "usage": {"input_tokens": 451, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.024149414883986524, "sagan": 0.6939311097058201, "socrates": 0.19740103402414771, "ibn_sina": 0.03407507211902187, "darwin": 0.020004439941672565, "hypatia": 0.03043892932535118}, "confidence": 0.4282942768796353}}, "usage": {"input_tokens": 451, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.5050608983752651, "ibn_sina": 0.020522407004047442, "hypatia": 0.031968449401848544, "einstein": 0.025089140137675197, "sagan": 0.39163374598956163, "darwin": 0.025725359091602014}, "confidence": 0.2934374944781014}}, "usage": {"input_tokens": 451, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:20Z"}
{"id": "c-hypatia-01-starter-01", "uid": "c050", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "prompt_sha256_by_order": ["49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7"], "choice": "sagan", "probs": {"darwin": 0.32348328575872665, "einstein": 0.019259632067121373, "hypatia": 0.05014849413512555, "ibn_sina": 0.017296053269095613, "sagan": 0.45918827190338185, "socrates": 0.130624262866549}, "confidence": 0.20289155394204927, "seconds": 0.19153076600014174, "seconds_all_orders": 1.1491845960008504, "seconds_by_order": [0.191114336001192, 0.191252374999749, 0.19212021800012735, 0.19178792999991856, 0.19164155200087407, 0.19126818499898945], "argmax_stable": true, "spread": {"darwin": 0.25934925998349073, "einstein": 0.005747255379789002, "hypatia": 0.014042761119057154, "ibn_sina": 0.007294492944507862, "sagan": 0.2763951943280284, "socrates": 0.04834805010731387}, "spread_2dp": {"darwin": 0.26, "einstein": 0.01, "hypatia": 0.010000000000000002, "ibn_sina": 0.01, "sagan": 0.28, "socrates": 0.04999999999999999}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.32348328575872665, "einstein": 0.019259632067121373, "hypatia": 0.05014849413512555, "ibn_sina": 0.017296053269095613, "sagan": 0.45918827190338185, "socrates": 0.130624262866549}, {"ibn_sina": 0.014113840885598875, "hypatia": 0.04738003918767404, "sagan": 0.39137393492439165, "socrates": 0.14744616473896138, "darwin": 0.3854898452371236, "einstein": 0.014196175026250524}, {"hypatia": 0.0436190672435184, "darwin": 0.3452516271499883, "sagan": 0.4353629952059092, "ibn_sina": 0.016134750709358076, "einstein": 0.019943430406039526, "socrates": 0.13968812928518648}, {"ibn_sina": 0.013742587902521088, "einstein": 0.016278213297220907, "darwin": 0.12614058525363286, "sagan": 0.6648029146508815, "hypatia": 0.04972299195763083, "socrates": 0.12931270693811286}, {"einstein": 0.017300940612370565, "sagan": 0.66776912925242, "socrates": 0.09909811463164751, "ibn_sina": 0.02103708084702895, "darwin": 0.15868900164046457, "hypatia": 0.0361057330160684}, {"socrates": 0.10919638400371404, "ibn_sina": 0.016007374391125845, "hypatia": 0.04796860633303184, "einstein": 0.01568693526117069, "sagan": 0.4445249358584236, "darwin": 0.36661576415253383}], "prompt_tokens": 559, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.32348328575872665, "einstein": 0.019259632067121373, "hypatia": 0.05014849413512555, "ibn_sina": 0.017296053269095613, "sagan": 0.45918827190338185, "socrates": 0.130624262866549}, "confidence": 0.20289155394204927}}, "usage": {"input_tokens": 559, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.014113840885598875, "hypatia": 0.04738003918767404, "sagan": 0.39137393492439165, "socrates": 0.14744616473896138, "darwin": 0.3854898452371236, "einstein": 0.014196175026250524}, "confidence": 0.19139433906246142}}, "usage": {"input_tokens": 559, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.0436190672435184, "darwin": 0.3452516271499883, "sagan": 0.4353629952059092, "ibn_sina": 0.016134750709358076, "einstein": 0.019943430406039526, "socrates": 0.13968812928518648}, "confidence": 0.19697570887789392}}, "usage": {"input_tokens": 559, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.013742587902521088, "einstein": 0.016278213297220907, "darwin": 0.12614058525363286, "sagan": 0.6648029146508815, "hypatia": 0.04972299195763083, "socrates": 0.12931270693811286}, "confidence": 0.37302682435807055}}, "usage": {"input_tokens": 559, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.017300940612370565, "sagan": 0.66776912925242, "socrates": 0.09909811463164751, "ibn_sina": 0.02103708084702895, "darwin": 0.15868900164046457, "hypatia": 0.0361057330160684}, "confidence": 0.3795565009851935}}, "usage": {"input_tokens": 559, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.10919638400371404, "ibn_sina": 0.016007374391125845, "hypatia": 0.04796860633303184, "einstein": 0.01568693526117069, "sagan": 0.4445249358584236, "darwin": 0.36661576415253383}, "confidence": 0.2160840286850653}}, "usage": {"input_tokens": 559, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:22Z"}
{"id": "c-hypatia-01-starter-02", "uid": "c051", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "prompt_sha256_by_order": ["d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd"], "choice": "ibn_sina", "probs": {"darwin": 0.005053853190170901, "einstein": 0.007357786632597377, "hypatia": 0.15100059274362862, "ibn_sina": 0.8300193222304699, "sagan": 0.0036318811224948762, "socrates": 0.0029365640806383888}, "confidence": 0.6542016960521705, "seconds": 0.18728960116671564, "seconds_all_orders": 1.1237376070002938, "seconds_by_order": [0.18708363000041572, 0.1874735959991085, 0.18754782600080944, 0.1877979040000355, 0.1869678599996405, 0.18686679100028414], "argmax_stable": false, "spread": {"darwin": 0.0030003187484731993, "einstein": 0.0037148771261473323, "hypatia": 0.44469359843789424, "ibn_sina": 0.44622689226806495, "sagan": 0.0031674558604237388, "socrates": 0.0018418482592167818}, "spread_2dp": {"darwin": 0.01, "einstein": 0.0, "hypatia": 0.43999999999999995, "ibn_sina": 0.43999999999999995, "sagan": 0.01, "socrates": 0.0}, "choices_by_order": ["ibn_sina", "ibn_sina", "hypatia", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.005053853190170901, "einstein": 0.007357786632597377, "hypatia": 0.15100059274362862, "ibn_sina": 0.8300193222304699, "sagan": 0.0036318811224948762, "socrates": 0.0029365640806383888}, {"ibn_sina": 0.5188728778208205, "hypatia": 0.4679414271069618, "sagan": 0.0037302807268274393, "socrates": 0.0018755615354426498, "darwin": 0.002053534441697702, "einstein": 0.005526318368250019}, {"hypatia": 0.583753870995111, "darwin": 0.004594636105197049, "sagan": 0.006799336982918615, "ibn_sina": 0.39641064315438895, "einstein": 0.005432000245696527, "socrates": 0.0030095125166878216}, {"ibn_sina": 0.5853490314845272, "einstein": 0.009146877371843859, "darwin": 0.004336051221044868, "sagan": 0.00576426684408394, "hypatia": 0.39168636328384066, "socrates": 0.0037174097946594317}, {"einstein": 0.006056101549642436, "sagan": 0.003671781050782032, "socrates": 0.0032587689062406333, "ibn_sina": 0.5508165381043193, "darwin": 0.004742779255043518, "hypatia": 0.4314540311339722}, {"socrates": 0.003247319495916679, "ibn_sina": 0.8426375354224539, "hypatia": 0.13906027255721679, "einstein": 0.006846588960568027, "sagan": 0.0048663448360184835, "darwin": 0.0033419387278261466}], "prompt_tokens": 513, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.005053853190170901, "einstein": 0.007357786632597377, "hypatia": 0.15100059274362862, "ibn_sina": 0.8300193222304699, "sagan": 0.0036318811224948762, "socrates": 0.0029365640806383888}, "confidence": 0.6542016960521705}}, "usage": {"input_tokens": 513, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.5188728778208205, "hypatia": 0.4679414271069618, "sagan": 0.0037302807268274393, "socrates": 0.0018755615354426498, "darwin": 0.002053534441697702, "einstein": 0.005526318368250019}, "confidence": 0.3859005189576938}}, "usage": {"input_tokens": 513, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.583753870995111, "darwin": 0.004594636105197049, "sagan": 0.006799336982918615, "ibn_sina": 0.39641064315438895, "einstein": 0.005432000245696527, "socrates": 0.0030095125166878216}, "confidence": 0.3976190624373482}}, "usage": {"input_tokens": 513, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.5853490314845272, "einstein": 0.009146877371843859, "darwin": 0.004336051221044868, "sagan": 0.00576426684408394, "hypatia": 0.39168636328384066, "socrates": 0.0037174097946594317}, "confidence": 0.3954414501472877}}, "usage": {"input_tokens": 513, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.006056101549642436, "sagan": 0.003671781050782032, "socrates": 0.0032587689062406333, "ibn_sina": 0.5508165381043193, "darwin": 0.004742779255043518, "hypatia": 0.4314540311339722}, "confidence": 0.387561653803471}}, "usage": {"input_tokens": 513, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.003247319495916679, "ibn_sina": 0.8426375354224539, "hypatia": 0.13906027255721679, "einstein": 0.006846588960568027, "sagan": 0.0048663448360184835, "darwin": 0.0033419387278261466}, "confidence": 0.6753616554848101}}, "usage": {"input_tokens": 513, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:23Z"}
{"id": "c-hypatia-01-starter-02", "uid": "c052", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "prompt_sha256_by_order": ["44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d"], "choice": "sagan", "probs": {"darwin": 0.04518400340722091, "einstein": 0.09372693594365758, "hypatia": 0.18159285257377905, "ibn_sina": 0.04041896671307627, "sagan": 0.6159047664592339, "socrates": 0.023172474903032377}, "confidence": 0.31037396152199237, "seconds": 0.18905249316685513, "seconds_all_orders": 1.1343149590011308, "seconds_by_order": [0.18911129299885943, 0.1889896040011081, 0.18955060900043463, 0.18885516400041524, 0.18896788399979414, 0.18884040500051924], "argmax_stable": true, "spread": {"darwin": 0.036105319282314904, "einstein": 0.0416248922856505, "hypatia": 0.19976001022658182, "ibn_sina": 0.054470219125093186, "sagan": 0.22767207145295765, "socrates": 0.004084240976282471}, "spread_2dp": {"darwin": 0.04000000000000001, "einstein": 0.05000000000000002, "hypatia": 0.19, "ibn_sina": 0.06, "sagan": 0.23000000000000004, "socrates": 0.0}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.04518400340722091, "einstein": 0.09372693594365758, "hypatia": 0.18159285257377905, "ibn_sina": 0.04041896671307627, "sagan": 0.6159047664592339, "socrates": 0.023172474903032377}, {"ibn_sina": 0.02789976261489771, "hypatia": 0.157082335670154, "sagan": 0.6174994071922304, "socrates": 0.02402493411036279, "darwin": 0.045097346317557574, "einstein": 0.12839621409479743}, {"hypatia": 0.18988012871986776, "darwin": 0.04398590316997168, "sagan": 0.5735874885536008, "ibn_sina": 0.03471896065937881, "einstein": 0.13535182822930808, "socrates": 0.022475690667872916}, {"ibn_sina": 0.02464294169165385, "einstein": 0.11030049600462727, "darwin": 0.04910311236818714, "sagan": 0.670860645885626, "hypatia": 0.12515211091582543, "socrates": 0.01994069313408032}, {"einstein": 0.12530734605997843, "sagan": 0.5856179713544345, "socrates": 0.020167532276582877, "ibn_sina": 0.07911316081674703, "darwin": 0.032883136497778284, "hypatia": 0.1569108529944789}, {"socrates": 0.023055536209269985, "ibn_sina": 0.03507446321161875, "hypatia": 0.32491212114240725, "einstein": 0.10478084922394248, "sagan": 0.44318857443266835, "darwin": 0.06898845578009319}], "prompt_tokens": 453, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.04518400340722091, "einstein": 0.09372693594365758, "hypatia": 0.18159285257377905, "ibn_sina": 0.04041896671307627, "sagan": 0.6159047664592339, "socrates": 0.023172474903032377}, "confidence": 0.31037396152199237}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.02789976261489771, "hypatia": 0.157082335670154, "sagan": 0.6174994071922304, "socrates": 0.02402493411036279, "darwin": 0.045097346317557574, "einstein": 0.12839621409479743}, "confidence": 0.3110263968569015}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.18988012871986776, "darwin": 0.04398590316997168, "sagan": 0.5735874885536008, "ibn_sina": 0.03471896065937881, "einstein": 0.13535182822930808, "socrates": 0.022475690667872916}, "confidence": 0.2644270123490323}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.02464294169165385, "einstein": 0.11030049600462727, "darwin": 0.04910311236818714, "sagan": 0.670860645885626, "hypatia": 0.12515211091582543, "socrates": 0.01994069313408032}, "confidence": 0.3775591335349141}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.12530734605997843, "sagan": 0.5856179713544345, "socrates": 0.020167532276582877, "ibn_sina": 0.07911316081674703, "darwin": 0.032883136497778284, "hypatia": 0.1569108529944789}, "confidence": 0.2692219328509702}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.023055536209269985, "ibn_sina": 0.03507446321161875, "hypatia": 0.32491212114240725, "einstein": 0.10478084922394248, "sagan": 0.44318857443266835, "darwin": 0.06898845578009319}, "confidence": 0.18338104970489408}}, "usage": {"input_tokens": 453, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:24Z"}
{"id": "c-hypatia-01-starter-02", "uid": "c053", "task": "kperm", "label": "hypatia", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "prompt_sha256_by_order": ["a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131"], "choice": "hypatia", "probs": {"darwin": 0.01408944867113416, "einstein": 0.027015015884682695, "hypatia": 0.5951454612011706, "ibn_sina": 0.029061906359040737, "sagan": 0.2504005833866759, "socrates": 0.08428758449729588}, "confidence": 0.31093106451694996, "seconds": 0.18846334150020994, "seconds_all_orders": 1.1307800490012596, "seconds_by_order": [0.18833435899978213, 0.18912949200057483, 0.18857370700061438, 0.1880341119995137, 0.18842520900034287, 0.1882831700004317], "argmax_stable": false, "spread": {"darwin": 0.005607042035236275, "einstein": 0.006847427921718489, "hypatia": 0.19330381859040857, "ibn_sina": 0.027148784530524397, "sagan": 0.2253735849346824, "socrates": 0.08018575477681968}, "spread_2dp": {"darwin": 0.01, "einstein": 0.009999999999999998, "hypatia": 0.19999999999999996, "ibn_sina": 0.03, "sagan": 0.22999999999999998, "socrates": 0.07999999999999999}, "choices_by_order": ["hypatia", "hypatia", "hypatia", "hypatia", "sagan", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.01408944867113416, "einstein": 0.027015015884682695, "hypatia": 0.5951454612011706, "ibn_sina": 0.029061906359040737, "sagan": 0.2504005833866759, "socrates": 0.08428758449729588}, {"ibn_sina": 0.026223206745605538, "hypatia": 0.5354461963023898, "sagan": 0.3275067253292629, "socrates": 0.06998404077851617, "darwin": 0.01236434481144296, "einstein": 0.028475486032782726}, {"hypatia": 0.4352306145580256, "darwin": 0.01612948615413556, "sagan": 0.35441684416151326, "ibn_sina": 0.029203725063902675, "einstein": 0.023784776248110373, "socrates": 0.1412345538143125}, {"ibn_sina": 0.03927774468109412, "einstein": 0.021628058111064236, "darwin": 0.017152989797338765, "sagan": 0.34231263541040086, "hypatia": 0.4840089683066767, "socrates": 0.09561960369342527}, {"einstein": 0.022802410871826023, "sagan": 0.4757741683213583, "socrates": 0.06736441424012128, "ibn_sina": 0.0206714161938298, "darwin": 0.01154594776210249, "hypatia": 0.40184164261076205}, {"socrates": 0.14755016901694096, "ibn_sina": 0.012128960150569724, "hypatia": 0.5351938955991367, "einstein": 0.022654403933378604, "sagan": 0.2653322080093301, "darwin": 0.017140363290643983}], "prompt_tokens": 525, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"darwin": 0.01408944867113416, "einstein": 0.027015015884682695, "hypatia": 0.5951454612011706, "ibn_sina": 0.029061906359040737, "sagan": 0.2504005833866759, "socrates": 0.08428758449729588}, "confidence": 0.31093106451694996}}, "usage": {"input_tokens": 525, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.026223206745605538, "hypatia": 0.5354461963023898, "sagan": 0.3275067253292629, "socrates": 0.06998404077851617, "darwin": 0.01236434481144296, "einstein": 0.028475486032782726}, "confidence": 0.28061492456047094}}, "usage": {"input_tokens": 525, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.4352306145580256, "darwin": 0.01612948615413556, "sagan": 0.35441684416151326, "ibn_sina": 0.029203725063902675, "einstein": 0.023784776248110373, "socrates": 0.1412345538143125}, "confidence": 0.20399550391296178}}, "usage": {"input_tokens": 525, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.03927774468109412, "einstein": 0.021628058111064236, "darwin": 0.017152989797338765, "sagan": 0.34231263541040086, "hypatia": 0.4840089683066767, "socrates": 0.09561960369342527}, "confidence": 0.23546856346872608}}, "usage": {"input_tokens": 525, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.022802410871826023, "sagan": 0.4757741683213583, "socrates": 0.06736441424012128, "ibn_sina": 0.0206714161938298, "darwin": 0.01154594776210249, "hypatia": 0.40184164261076205}, "confidence": 0.27214755469919394}}, "usage": {"input_tokens": 525, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"socrates": 0.14755016901694096, "ibn_sina": 0.012128960150569724, "hypatia": 0.5351938955991367, "einstein": 0.022654403933378604, "sagan": 0.2653322080093301, "darwin": 0.017140363290643983}, "confidence": 0.2554706375395119}}, "usage": {"input_tokens": 525, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:25Z"}
{"id": "c-ibn_sina-01-dessert-01", "uid": "c054", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "prompt_sha256_by_order": ["2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d"], "choice": "sagan", "probs": {"darwin": 0.04115436406130027, "einstein": 0.13615477508257343, "hypatia": 0.265796805089607, "ibn_sina": 0.054417912825330914, "sagan": 0.4593896508881357, "socrates": 0.04308649205305264}, "confidence": 0.16808362291809403, "seconds": 0.18913779599946187, "seconds_all_orders": 1.1348267759967712, "seconds_by_order": [0.18944547999853967, 0.18926481199923728, 0.18924015199991118, 0.18882434499937517, 0.188699626000016, 0.18935236099969188], "argmax_stable": true, "spread": {"darwin": 0.029256347141146916, "einstein": 0.1104169666863534, "hypatia": 0.1041332434308713, "ibn_sina": 0.03012110123228138, "sagan": 0.09366178318970703, "socrates": 0.029543637809275904}, "spread_2dp": {"darwin": 0.030000000000000002, "einstein": 0.10999999999999999, "hypatia": 0.11000000000000001, "ibn_sina": 0.030000000000000002, "sagan": 0.09000000000000002, "socrates": 0.03}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.04115436406130027, "einstein": 0.13615477508257343, "hypatia": 0.265796805089607, "ibn_sina": 0.054417912825330914, "sagan": 0.4593896508881357, "socrates": 0.04308649205305264}, {"ibn_sina": 0.036015128676763444, "hypatia": 0.16166356165873572, "sagan": 0.4980787498338599, "socrates": 0.03395929482060043, "darwin": 0.031143963432577675, "einstein": 0.23913930157746283}, {"hypatia": 0.2585730912439799, "darwin": 0.03952794561476028, "sagan": 0.4598283851742807, "ibn_sina": 0.0366512876113109, "einstein": 0.14908843894099047, "socrates": 0.05633085141467769}, {"ibn_sina": 0.050104545815899904, "einstein": 0.12872233489110943, "darwin": 0.04735471361181111, "sagan": 0.47834842740391836, "hypatia": 0.23661279699166649, "socrates": 0.058857181285594695}, {"einstein": 0.1552073862029695, "sagan": 0.5530514340778427, "socrates": 0.02931354347631879, "ibn_sina": 0.024296811593049535, "darwin": 0.02174657754199001, "hypatia": 0.21638424710782939}, {"socrates": 0.04417718586189195, "ibn_sina": 0.03582643198993174, "hypatia": 0.19428108975814334, "einstein": 0.13264675272106646, "sagan": 0.5420656149858296, "darwin": 0.05100292468313693}], "prompt_tokens": 529, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.04115436406130027, "einstein": 0.13615477508257343, "hypatia": 0.265796805089607, "ibn_sina": 0.054417912825330914, "sagan": 0.4593896508881357, "socrates": 0.04308649205305264}, "confidence": 0.16808362291809403}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.036015128676763444, "hypatia": 0.16166356165873572, "sagan": 0.4980787498338599, "socrates": 0.03395929482060043, "darwin": 0.031143963432577675, "einstein": 0.23913930157746283}, "confidence": 0.20179050810377372}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.2585730912439799, "darwin": 0.03952794561476028, "sagan": 0.4598283851742807, "ibn_sina": 0.0366512876113109, "einstein": 0.14908843894099047, "socrates": 0.05633085141467769}, "confidence": 0.16793018817101038}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.050104545815899904, "einstein": 0.12872233489110943, "darwin": 0.04735471361181111, "sagan": 0.47834842740391836, "hypatia": 0.23661279699166649, "socrates": 0.058857181285594695}, "confidence": 0.17150725048167983}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.1552073862029695, "sagan": 0.5530514340778427, "socrates": 0.02931354347631879, "ibn_sina": 0.024296811593049535, "darwin": 0.02174657754199001, "hypatia": 0.21638424710782939}, "confidence": 0.2544398756601452}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.04417718586189195, "ibn_sina": 0.03582643198993174, "hypatia": 0.19428108975814334, "einstein": 0.13264675272106646, "sagan": 0.5420656149858296, "darwin": 0.05100292468313693}, "confidence": 0.22601426692127227}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:26Z"}
{"id": "c-ibn_sina-01-dessert-02", "uid": "c055", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "prompt_sha256_by_order": ["a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e"], "choice": "ibn_sina", "probs": {"darwin": 0.011260686289451906, "einstein": 0.01786333312480521, "hypatia": 0.005052682352387549, "ibn_sina": 0.9491087176197535, "sagan": 0.004149742877911354, "socrates": 0.012564837735690464}, "confidence": 0.8817446616400975, "seconds": 0.1912914193329319, "seconds_all_orders": 1.1477485159975913, "seconds_by_order": [0.19166732099984074, 0.19108442600008857, 0.19105046599906927, 0.19128772499971092, 0.19121762499889883, 0.19144095299998298], "argmax_stable": true, "spread": {"darwin": 0.006126959966602584, "einstein": 0.04286507418645079, "hypatia": 0.001363584030686568, "ibn_sina": 0.0347915637496613, "sagan": 0.0016335122534781692, "socrates": 0.013422011111275353}, "spread_2dp": {"darwin": 0.0, "einstein": 0.039999999999999994, "hypatia": 0.01, "ibn_sina": 0.029999999999999916, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.011260686289451906, "einstein": 0.01786333312480521, "hypatia": 0.005052682352387549, "ibn_sina": 0.9491087176197535, "sagan": 0.004149742877911354, "socrates": 0.012564837735690464}, {"ibn_sina": 0.9470172859757896, "hypatia": 0.004509276197730818, "sagan": 0.004132310873386074, "socrates": 0.007496159643349565, "darwin": 0.006990608261505649, "einstein": 0.029854359048238334}, {"hypatia": 0.0056837471798930235, "darwin": 0.005133726322849322, "sagan": 0.004696996247465477, "ibn_sina": 0.9449570330467512, "einstein": 0.023280201232025077, "socrates": 0.016248295971016027}, {"ibn_sina": 0.9442149509019135, "einstein": 0.021813717937896833, "darwin": 0.0063033937160013815, "sagan": 0.0031883771485573366, "hypatia": 0.0043201631492064555, "socrates": 0.020159397146424507}, {"einstein": 0.05880444639276188, "sagan": 0.003648632228874093, "socrates": 0.006737386035149154, "ibn_sina": 0.9184688826002538, "darwin": 0.007059109914979237, "hypatia": 0.005281542827981775}, {"socrates": 0.012232916023627024, "ibn_sina": 0.9532604463499151, "hypatia": 0.0047624122338844055, "einstein": 0.015939372206311097, "sagan": 0.003063483993987308, "darwin": 0.010741369192274976}], "prompt_tokens": 551, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.011260686289451906, "einstein": 0.01786333312480521, "hypatia": 0.005052682352387549, "ibn_sina": 0.9491087176197535, "sagan": 0.004149742877911354, "socrates": 0.012564837735690464}, "confidence": 0.8817446616400975}}, "usage": {"input_tokens": 551, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.9470172859757896, "hypatia": 0.004509276197730818, "sagan": 0.004132310873386074, "socrates": 0.007496159643349565, "darwin": 0.006990608261505649, "einstein": 0.029854359048238334}, "confidence": 0.8774505919232534}}, "usage": {"input_tokens": 551, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.0056837471798930235, "darwin": 0.005133726322849322, "sagan": 0.004696996247465477, "ibn_sina": 0.9449570330467512, "einstein": 0.023280201232025077, "socrates": 0.016248295971016027}, "confidence": 0.8725965893171146}}, "usage": {"input_tokens": 551, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.9442149509019135, "einstein": 0.021813717937896833, "darwin": 0.0063033937160013815, "sagan": 0.0031883771485573366, "hypatia": 0.0043201631492064555, "socrates": 0.020159397146424507}, "confidence": 0.8709912105053206}}, "usage": {"input_tokens": 551, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.05880444639276188, "sagan": 0.003648632228874093, "socrates": 0.006737386035149154, "ibn_sina": 0.9184688826002538, "darwin": 0.007059109914979237, "hypatia": 0.005281542827981775}, "confidence": 0.8166153782028183}}, "usage": {"input_tokens": 551, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.012232916023627024, "ibn_sina": 0.9532604463499151, "hypatia": 0.0047624122338844055, "einstein": 0.015939372206311097, "sagan": 0.003063483993987308, "darwin": 0.010741369192274976}, "confidence": 0.8911079546951243}}, "usage": {"input_tokens": 551, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:27Z"}
{"id": "c-ibn_sina-01-dessert-02", "uid": "c056", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "prompt_sha256_by_order": ["ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595"], "choice": "hypatia", "probs": {"darwin": 0.05931848285694919, "einstein": 0.02127267467136473, "hypatia": 0.38324131241857373, "ibn_sina": 0.11151521984622942, "sagan": 0.06596838365644049, "socrates": 0.35868392655044246}, "confidence": 0.15554409244680895, "seconds": 0.22176350550004523, "seconds_all_orders": 1.3305810330002714, "seconds_by_order": [0.2213547480005218, 0.22182104300009087, 0.22210541099957481, 0.22201468200000818, 0.22178044399879582, 0.22150470500127994], "argmax_stable": false, "spread": {"darwin": 0.07501834832344044, "einstein": 0.008909848804715102, "hypatia": 0.44334468157534235, "ibn_sina": 0.1533117336346205, "sagan": 0.09180686261092026, "socrates": 0.29478846464190533}, "spread_2dp": {"darwin": 0.08, "einstein": 0.0, "hypatia": 0.44999999999999996, "ibn_sina": 0.15, "sagan": 0.09, "socrates": 0.3}, "choices_by_order": ["hypatia", "hypatia", "hypatia", "hypatia", "hypatia", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.05931848285694919, "einstein": 0.02127267467136473, "hypatia": 0.38324131241857373, "ibn_sina": 0.11151521984622942, "sagan": 0.06596838365644049, "socrates": 0.35868392655044246}, {"ibn_sina": 0.15483032829747898, "hypatia": 0.5945094569354327, "sagan": 0.07371486571347587, "socrates": 0.10484524854677092, "darwin": 0.050163211796936315, "einstein": 0.021936888709905303}, {"hypatia": 0.6126690114366998, "darwin": 0.04704577196376619, "sagan": 0.026627622814592966, "ibn_sina": 0.0650574027954973, "einstein": 0.015779519190620213, "socrates": 0.23282067179882354}, {"ibn_sina": 0.2056670303922261, "einstein": 0.01914188024465876, "darwin": 0.036171653143135654, "sagan": 0.030330759753941804, "hypatia": 0.4103492317038555, "socrates": 0.29833944476218216}, {"einstein": 0.021057297403547826, "sagan": 0.08223980454102428, "socrates": 0.06389546190853715, "ibn_sina": 0.09403187397498937, "darwin": 0.03044538150905017, "hypatia": 0.7083301806628512}, {"socrates": 0.2680577812290342, "ibn_sina": 0.2183691364301178, "hypatia": 0.2649854990875089, "einstein": 0.024689367995335314, "sagan": 0.11843448542551323, "darwin": 0.10546372983249061}], "prompt_tokens": 591, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"darwin": 0.05931848285694919, "einstein": 0.02127267467136473, "hypatia": 0.38324131241857373, "ibn_sina": 0.11151521984622942, "sagan": 0.06596838365644049, "socrates": 0.35868392655044246}, "confidence": 0.15554409244680895}}, "usage": {"input_tokens": 591, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.15483032829747898, "hypatia": 0.5945094569354327, "sagan": 0.07371486571347587, "socrates": 0.10484524854677092, "darwin": 0.050163211796936315, "einstein": 0.021936888709905303}, "confidence": 0.27620548890449587}}, "usage": {"input_tokens": 591, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.6126690114366998, "darwin": 0.04704577196376619, "sagan": 0.026627622814592966, "ibn_sina": 0.0650574027954973, "einstein": 0.015779519190620213, "socrates": 0.23282067179882354}, "confidence": 0.32436709195898106}}, "usage": {"input_tokens": 591, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.2056670303922261, "einstein": 0.01914188024465876, "darwin": 0.036171653143135654, "sagan": 0.030330759753941804, "hypatia": 0.4103492317038555, "socrates": 0.29833944476218216}, "confidence": 0.1627439184507616}}, "usage": {"input_tokens": 591, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.021057297403547826, "sagan": 0.08223980454102428, "socrates": 0.06389546190853715, "ibn_sina": 0.09403187397498937, "darwin": 0.03044538150905017, "hypatia": 0.7083301806628512}, "confidence": 0.42734798163248805}}, "usage": {"input_tokens": 591, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.2680577812290342, "ibn_sina": 0.2183691364301178, "hypatia": 0.2649854990875089, "einstein": 0.024689367995335314, "sagan": 0.11843448542551323, "darwin": 0.10546372983249061}, "confidence": 0.05861951090742408}}, "usage": {"input_tokens": 591, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:29Z"}
{"id": "c-ibn_sina-01-dessert-02", "uid": "c057", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "prompt_sha256_by_order": ["fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e"], "choice": "ibn_sina", "probs": {"darwin": 0.049525084487437164, "einstein": 0.039959939101446906, "hypatia": 0.006892669146923679, "ibn_sina": 0.8871913646370876, "sagan": 0.0051387026701651415, "socrates": 0.011292239956939534}, "confidence": 0.7496313736593319, "seconds": 0.22177718883297834, "seconds_all_orders": 1.33066313299787, "seconds_by_order": [0.22199208199890563, 0.2216008059986052, 0.22203639100007422, 0.2211610690010275, 0.22176989400031744, 0.22210289099893998], "argmax_stable": true, "spread": {"darwin": 0.02722096434150135, "einstein": 0.04060487797966871, "hypatia": 0.005470045160530367, "ibn_sina": 0.037008350130940615, "sagan": 0.0023805342771087996, "socrates": 0.02907531481288576}, "spread_2dp": {"darwin": 0.030000000000000002, "einstein": 0.039999999999999994, "hypatia": 0.0, "ibn_sina": 0.030000000000000027, "sagan": 0.01, "socrates": 0.03}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.049525084487437164, "einstein": 0.039959939101446906, "hypatia": 0.006892669146923679, "ibn_sina": 0.8871913646370876, "sagan": 0.0051387026701651415, "socrates": 0.011292239956939534}, {"ibn_sina": 0.8766374923127926, "hypatia": 0.006931706241621818, "sagan": 0.004786621134182352, "socrates": 0.006170549917343049, "darwin": 0.04772302347248316, "einstein": 0.05775060692157703}, {"hypatia": 0.010763253882219598, "darwin": 0.022304120145935816, "sagan": 0.006155079583289855, "ibn_sina": 0.9136458424437333, "einstein": 0.03719579064336816, "socrates": 0.00993591330145327}, {"ibn_sina": 0.8850946718082258, "einstein": 0.03475395881719786, "darwin": 0.03479840664105717, "sagan": 0.004521753220712809, "hypatia": 0.005585344782577572, "socrates": 0.03524586473022881}, {"einstein": 0.0466251027497124, "sagan": 0.006902287497821609, "socrates": 0.006701741150116132, "ibn_sina": 0.9005143241728726, "darwin": 0.02820115448636934, "hypatia": 0.01105538994310794}, {"socrates": 0.012789552541591977, "ibn_sina": 0.9129061252356256, "hypatia": 0.006194200469477953, "einstein": 0.01714572894190832, "sagan": 0.004777205328383119, "darwin": 0.04618718748301297}], "prompt_tokens": 585, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.049525084487437164, "einstein": 0.039959939101446906, "hypatia": 0.006892669146923679, "ibn_sina": 0.8871913646370876, "sagan": 0.0051387026701651415, "socrates": 0.011292239956939534}, "confidence": 0.7496313736593319}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.8766374923127926, "hypatia": 0.006931706241621818, "sagan": 0.004786621134182352, "socrates": 0.006170549917343049, "darwin": 0.04772302347248316, "einstein": 0.05775060692157703}, "confidence": 0.7290579381726545}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.010763253882219598, "darwin": 0.022304120145935816, "sagan": 0.006155079583289855, "ibn_sina": 0.9136458424437333, "einstein": 0.03719579064336816, "socrates": 0.00993591330145327}, "confidence": 0.8042586172524789}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.8850946718082258, "einstein": 0.03475395881719786, "darwin": 0.03479840664105717, "sagan": 0.004521753220712809, "hypatia": 0.005585344782577572, "socrates": 0.03524586473022881}, "confidence": 0.7445263097567747}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.0466251027497124, "sagan": 0.006902287497821609, "socrates": 0.006701741150116132, "ibn_sina": 0.9005143241728726, "darwin": 0.02820115448636934, "hypatia": 0.01105538994310794}, "confidence": 0.776932035898301}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.012789552541591977, "ibn_sina": 0.9129061252356256, "hypatia": 0.006194200469477953, "einstein": 0.01714572894190832, "sagan": 0.004777205328383119, "darwin": 0.04618718748301297}, "confidence": 0.8032595059188207}}, "usage": {"input_tokens": 585, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:30Z"}
{"id": "c-ibn_sina-01-dessert-03", "uid": "c058", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "prompt_sha256_by_order": ["a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c"], "choice": "ibn_sina", "probs": {"darwin": 0.006890309423275013, "einstein": 0.008930778077562375, "hypatia": 0.24763966021641268, "ibn_sina": 0.6821989075862599, "sagan": 0.01157535855272174, "socrates": 0.04276498614376828}, "confidence": 0.434572982740608, "seconds": 0.22397215383292254, "seconds_all_orders": 1.3438329229975352, "seconds_by_order": [0.22418800400009786, 0.22339195999848016, 0.22395810499983781, 0.22402145500018378, 0.22437114300009853, 0.22390225599883706], "argmax_stable": false, "spread": {"darwin": 0.0026730130788907144, "einstein": 0.0032863537428356217, "hypatia": 0.2426040227749006, "ibn_sina": 0.25725343836634884, "sagan": 0.0709920238435513, "socrates": 0.03316294632978506}, "spread_2dp": {"darwin": 0.01, "einstein": 0.0, "hypatia": 0.24, "ibn_sina": 0.26000000000000006, "sagan": 0.07, "socrates": 0.03}, "choices_by_order": ["ibn_sina", "ibn_sina", "hypatia", "hypatia", "hypatia", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.006890309423275013, "einstein": 0.008930778077562375, "hypatia": 0.24763966021641268, "ibn_sina": 0.6821989075862599, "sagan": 0.01157535855272174, "socrates": 0.04276498614376828}, {"ibn_sina": 0.5606163367199591, "hypatia": 0.3817728471111427, "sagan": 0.028153046084044668, "socrates": 0.01609727492464523, "darwin": 0.004785622541762556, "einstein": 0.008574872618445722}, {"hypatia": 0.47388414220544456, "darwin": 0.005826321126489191, "sagan": 0.018765249499347172, "ibn_sina": 0.4652840551361056, "einstein": 0.007582214375063453, "socrates": 0.028658017657550022}, {"ibn_sina": 0.42494546921991105, "einstein": 0.009430469047430435, "darwin": 0.00584580039505821, "sagan": 0.08256738239627304, "hypatia": 0.4379128627430733, "socrates": 0.039298016198253974}, {"einstein": 0.010868568117899075, "sagan": 0.04400558721834171, "socrates": 0.009602039813983217, "ibn_sina": 0.43935406673281796, "darwin": 0.0059260551256448, "hypatia": 0.49024368299131327}, {"socrates": 0.023202260233159713, "ibn_sina": 0.6181738401143564, "hypatia": 0.29026512275893146, "einstein": 0.010728176447623049, "sagan": 0.05017196482527605, "darwin": 0.00745863562065327}], "prompt_tokens": 609, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.006890309423275013, "einstein": 0.008930778077562375, "hypatia": 0.24763966021641268, "ibn_sina": 0.6821989075862599, "sagan": 0.01157535855272174, "socrates": 0.04276498614376828}, "confidence": 0.434572982740608}}, "usage": {"input_tokens": 609, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.5606163367199591, "hypatia": 0.3817728471111427, "sagan": 0.028153046084044668, "socrates": 0.01609727492464523, "darwin": 0.004785622541762556, "einstein": 0.008574872618445722}, "confidence": 0.35342719681119916}}, "usage": {"input_tokens": 609, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.47388414220544456, "darwin": 0.005826321126489191, "sagan": 0.018765249499347172, "ibn_sina": 0.4652840551361056, "einstein": 0.007582214375063453, "socrates": 0.028658017657550022}, "confidence": 0.33078434170626425}}, "usage": {"input_tokens": 609, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.42494546921991105, "einstein": 0.009430469047430435, "darwin": 0.00584580039505821, "sagan": 0.08256738239627304, "hypatia": 0.4379128627430733, "socrates": 0.039298016198253974}, "confidence": 0.25699736920956184}}, "usage": {"input_tokens": 609, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.010868568117899075, "sagan": 0.04400558721834171, "socrates": 0.009602039813983217, "ibn_sina": 0.43935406673281796, "darwin": 0.0059260551256448, "hypatia": 0.49024368299131327}, "confidence": 0.32266335933383045}}, "usage": {"input_tokens": 609, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.023202260233159713, "ibn_sina": 0.6181738401143564, "hypatia": 0.29026512275893146, "einstein": 0.010728176447623049, "sagan": 0.05017196482527605, "darwin": 0.00745863562065327}, "confidence": 0.36354284084986765}}, "usage": {"input_tokens": 609, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:31Z"}
{"id": "c-ibn_sina-01-dessert-03", "uid": "c059", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "prompt_sha256_by_order": ["5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de"], "choice": "ibn_sina", "probs": {"darwin": 0.12139468515437955, "einstein": 0.004786949442278072, "hypatia": 0.009623795503443142, "ibn_sina": 0.8495358162734593, "sagan": 0.009491190327427837, "socrates": 0.00516756329901202}, "confidence": 0.6840161097331355, "seconds": 0.2535995446666372, "seconds_all_orders": 1.521597267999823, "seconds_by_order": [0.25321490399983304, 0.25353926199932175, 0.25370900000052643, 0.2543328449992259, 0.25303374599934614, 0.2537675110015698], "argmax_stable": true, "spread": {"darwin": 0.07439332717792159, "einstein": 0.002184441076318093, "hypatia": 0.03817154285320018, "ibn_sina": 0.05219071549043497, "sagan": 0.0060326683892850674, "socrates": 0.0021689893824447387}, "spread_2dp": {"darwin": 0.07, "einstein": 0.01, "hypatia": 0.04, "ibn_sina": 0.04999999999999993, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.12139468515437955, "einstein": 0.004786949442278072, "hypatia": 0.009623795503443142, "ibn_sina": 0.8495358162734593, "sagan": 0.009491190327427837, "socrates": 0.00516756329901202}, {"ibn_sina": 0.8158618368920476, "hypatia": 0.010741473076035634, "sagan": 0.009133506215037676, "socrates": 0.004127023663135378, "darwin": 0.15510486005767177, "einstein": 0.00503130009607192}, {"hypatia": 0.032805880427532005, "darwin": 0.08743955387078096, "sagan": 0.0073533755122177705, "ibn_sina": 0.861166617071681, "einstein": 0.006420833488765855, "socrates": 0.004813739629022568}, {"ibn_sina": 0.8133371010189299, "einstein": 0.006971390518596165, "darwin": 0.12766146995347027, "sagan": 0.013386043901502838, "hypatia": 0.03360456638203129, "socrates": 0.005039428225469435}, {"einstein": 0.006568414440279859, "sagan": 0.007932457121365866, "socrates": 0.00500178170701518, "ibn_sina": 0.8261974491814952, "darwin": 0.10799320855978008, "hypatia": 0.0463066889900638}, {"socrates": 0.0062960130455801165, "ibn_sina": 0.808975901581246, "hypatia": 0.008135146136863619, "einstein": 0.005529085408109689, "sagan": 0.009230972779498001, "darwin": 0.16183288104870255}], "prompt_tokens": 689, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.12139468515437955, "einstein": 0.004786949442278072, "hypatia": 0.009623795503443142, "ibn_sina": 0.8495358162734593, "sagan": 0.009491190327427837, "socrates": 0.00516756329901202}, "confidence": 0.6840161097331355}}, "usage": {"input_tokens": 689, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.8158618368920476, "hypatia": 0.010741473076035634, "sagan": 0.009133506215037676, "socrates": 0.004127023663135378, "darwin": 0.15510486005767177, "einstein": 0.00503130009607192}, "confidence": 0.6279150411939668}}, "usage": {"input_tokens": 689, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.032805880427532005, "darwin": 0.08743955387078096, "sagan": 0.0073533755122177705, "ibn_sina": 0.861166617071681, "einstein": 0.006420833488765855, "socrates": 0.004813739629022568}, "confidence": 0.7005379780645121}}, "usage": {"input_tokens": 689, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.8133371010189299, "einstein": 0.006971390518596165, "darwin": 0.12766146995347027, "sagan": 0.013386043901502838, "hypatia": 0.03360456638203129, "socrates": 0.005039428225469435}, "confidence": 0.6150365679762609}}, "usage": {"input_tokens": 689, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.006568414440279859, "sagan": 0.007932457121365866, "socrates": 0.00500178170701518, "ibn_sina": 0.8261974491814952, "darwin": 0.10799320855978008, "hypatia": 0.0463066889900638}, "confidence": 0.635848184006505}}, "usage": {"input_tokens": 689, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.0062960130455801165, "ibn_sina": 0.808975901581246, "hypatia": 0.008135146136863619, "einstein": 0.005529085408109689, "sagan": 0.009230972779498001, "darwin": 0.16183288104870255}, "confidence": 0.6170241913054693}}, "usage": {"input_tokens": 689, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:33Z"}
{"id": "c-ibn_sina-01-main-02", "uid": "c060", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "prompt_sha256_by_order": ["1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3"], "choice": "hypatia", "probs": {"darwin": 0.008646533980499855, "einstein": 0.03404191207062021, "hypatia": 0.43719128763102094, "ibn_sina": 0.10963111094410039, "sagan": 0.3704961586148961, "socrates": 0.03999299675886248}, "confidence": 0.2119067921593074, "seconds": 0.19064437666687203, "seconds_all_orders": 1.1438662600012321, "seconds_by_order": [0.19047618100012187, 0.19061393000083626, 0.19035367200012843, 0.19086855899877264, 0.19083744900126476, 0.1907164690001082], "argmax_stable": false, "spread": {"darwin": 0.006305618207541371, "einstein": 0.08619888326103775, "hypatia": 0.24915322359330383, "ibn_sina": 0.04809229424013631, "sagan": 0.3041727332515565, "socrates": 0.03861506549238891}, "spread_2dp": {"darwin": 0.0, "einstein": 0.09, "hypatia": 0.25000000000000006, "ibn_sina": 0.05, "sagan": 0.31000000000000005, "socrates": 0.03}, "choices_by_order": ["hypatia", "hypatia", "sagan", "hypatia", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.008646533980499855, "einstein": 0.03404191207062021, "hypatia": 0.43719128763102094, "ibn_sina": 0.10963111094410039, "sagan": 0.3704961586148961, "socrates": 0.03999299675886248}, {"ibn_sina": 0.06537802119711672, "hypatia": 0.5501749999993438, "sagan": 0.254336100482006, "socrates": 0.037323041474924545, "darwin": 0.008647302609672943, "einstein": 0.08414053423693599}, {"hypatia": 0.35385506939960415, "darwin": 0.00794846615533517, "sagan": 0.4416250020682202, "ibn_sina": 0.09344304203235715, "einstein": 0.040315492182495624, "socrates": 0.06281292816198772}, {"ibn_sina": 0.06923649945889479, "einstein": 0.12024079533165796, "darwin": 0.013357944534531027, "sagan": 0.2555073223260829, "hypatia": 0.4780005511412844, "socrates": 0.06365688720754885}, {"einstein": 0.04464756225067688, "sagan": 0.5541616716437181, "socrates": 0.025948026312433536, "ibn_sina": 0.06153881670396408, "darwin": 0.007052326326989656, "hypatia": 0.3066515967622177}, {"socrates": 0.025041821715159934, "ibn_sina": 0.06574470962992068, "hypatia": 0.30102177640604, "einstein": 0.03929553064746866, "sagan": 0.5585088337335625, "darwin": 0.01038732786784828}], "prompt_tokens": 541, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"darwin": 0.008646533980499855, "einstein": 0.03404191207062021, "hypatia": 0.43719128763102094, "ibn_sina": 0.10963111094410039, "sagan": 0.3704961586148961, "socrates": 0.03999299675886248}, "confidence": 0.2119067921593074}}, "usage": {"input_tokens": 541, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.06537802119711672, "hypatia": 0.5501749999993438, "sagan": 0.254336100482006, "socrates": 0.037323041474924545, "darwin": 0.008647302609672943, "einstein": 0.08414053423693599}, "confidence": 0.256241299668833}}, "usage": {"input_tokens": 541, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.35385506939960415, "darwin": 0.00794846615533517, "sagan": 0.4416250020682202, "ibn_sina": 0.09344304203235715, "einstein": 0.040315492182495624, "socrates": 0.06281292816198772}, "confidence": 0.20153396279708496}}, "usage": {"input_tokens": 541, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.06923649945889479, "einstein": 0.12024079533165796, "darwin": 0.013357944534531027, "sagan": 0.2555073223260829, "hypatia": 0.4780005511412844, "socrates": 0.06365688720754885}, "confidence": 0.18070083321288388}}, "usage": {"input_tokens": 541, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.04464756225067688, "sagan": 0.5541616716437181, "socrates": 0.025948026312433536, "ibn_sina": 0.06153881670396408, "darwin": 0.007052326326989656, "hypatia": 0.3066515967622177}, "confidence": 0.2891605915217957}}, "usage": {"input_tokens": 541, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.025041821715159934, "ibn_sina": 0.06574470962992068, "hypatia": 0.30102177640604, "einstein": 0.03929553064746866, "sagan": 0.5585088337335625, "darwin": 0.01038732786784828}, "confidence": 0.2909772666607706}}, "usage": {"input_tokens": 541, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:34Z"}
{"id": "c-ibn_sina-01-main-02", "uid": "c061", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "prompt_sha256_by_order": ["0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c"], "choice": "einstein", "probs": {"darwin": 0.024723785356501847, "einstein": 0.8962259787848946, "hypatia": 0.007730583789732648, "ibn_sina": 0.05681439024312944, "sagan": 0.008009658373675368, "socrates": 0.006495603452066025}, "confidence": 0.768671505960634, "seconds": 0.22408702799960642, "seconds_all_orders": 1.3445221679976385, "seconds_by_order": [0.22397089599871833, 0.2236285980015964, 0.22472678899976017, 0.2239575359999435, 0.22416345399869897, 0.22407489499892108], "argmax_stable": true, "spread": {"darwin": 0.03155929868291027, "einstein": 0.11726634962460725, "hypatia": 0.004482981642975915, "ibn_sina": 0.10200531587328891, "sagan": 0.006652385661864256, "socrates": 0.009629636890087254}, "spread_2dp": {"darwin": 0.039999999999999994, "einstein": 0.12, "hypatia": 0.0, "ibn_sina": 0.1, "sagan": 0.01, "socrates": 0.0}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.024723785356501847, "einstein": 0.8962259787848946, "hypatia": 0.007730583789732648, "ibn_sina": 0.05681439024312944, "sagan": 0.008009658373675368, "socrates": 0.006495603452066025}, {"ibn_sina": 0.15881970611641835, "hypatia": 0.0072848410911863625, "sagan": 0.0060134729177847065, "socrates": 0.005015153027976757, "darwin": 0.0439071976863465, "einstein": 0.7789596291602874}, {"hypatia": 0.009691534861627818, "darwin": 0.030637101456097174, "sagan": 0.006805620890557796, "ibn_sina": 0.06430642022744269, "einstein": 0.8798691582101658, "socrates": 0.008690164354108833}, {"ibn_sina": 0.08987791622619731, "einstein": 0.8634261930501008, "darwin": 0.03041080208987225, "sagan": 0.004579594545207363, "hypatia": 0.005751595161921066, "socrates": 0.005953898926701202}, {"einstein": 0.8063267472213125, "sagan": 0.007408266279394936, "socrates": 0.00690932659205863, "ibn_sina": 0.11283799906292491, "darwin": 0.056283084039412117, "hypatia": 0.010234576804896981}, {"socrates": 0.01464478991806401, "ibn_sina": 0.07170947944898275, "hypatia": 0.007875655856402156, "einstein": 0.8622036694919641, "sagan": 0.01123198020707162, "darwin": 0.03233442507751537}], "prompt_tokens": 611, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.024723785356501847, "einstein": 0.8962259787848946, "hypatia": 0.007730583789732648, "ibn_sina": 0.05681439024312944, "sagan": 0.008009658373675368, "socrates": 0.006495603452066025}, "confidence": 0.768671505960634}}, "usage": {"input_tokens": 611, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.15881970611641835, "hypatia": 0.0072848410911863625, "sagan": 0.0060134729177847065, "socrates": 0.005015153027976757, "darwin": 0.0439071976863465, "einstein": 0.7789596291602874}, "confidence": 0.560852832936708}}, "usage": {"input_tokens": 611, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.009691534861627818, "darwin": 0.030637101456097174, "sagan": 0.006805620890557796, "ibn_sina": 0.06430642022744269, "einstein": 0.8798691582101658, "socrates": 0.008690164354108833}, "confidence": 0.7353513334212962}}, "usage": {"input_tokens": 611, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.08987791622619731, "einstein": 0.8634261930501008, "darwin": 0.03041080208987225, "sagan": 0.004579594545207363, "hypatia": 0.005751595161921066, "socrates": 0.005953898926701202}, "confidence": 0.7055165799993246}}, "usage": {"input_tokens": 611, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.8063267472213125, "sagan": 0.007408266279394936, "socrates": 0.00690932659205863, "ibn_sina": 0.11283799906292491, "darwin": 0.056283084039412117, "hypatia": 0.010234576804896981}, "confidence": 0.5995244687579344}}, "usage": {"input_tokens": 611, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.01464478991806401, "ibn_sina": 0.07170947944898275, "hypatia": 0.007875655856402156, "einstein": 0.8622036694919641, "sagan": 0.01123198020707162, "darwin": 0.03233442507751537}, "confidence": 0.6999827024555565}}, "usage": {"input_tokens": 611, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:35Z"}
{"id": "c-ibn_sina-01-main-02", "uid": "c062", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "prompt_sha256_by_order": ["c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de"], "choice": "ibn_sina", "probs": {"darwin": 0.00723911479158295, "einstein": 0.004621542056406089, "hypatia": 0.1093134469094677, "ibn_sina": 0.8632335117805118, "sagan": 0.007406085461505925, "socrates": 0.008186299000525537}, "confidence": 0.7087805854759335, "seconds": 0.22664862850009135, "seconds_all_orders": 1.359891771000548, "seconds_by_order": [0.22672330199930002, 0.2263409960014542, 0.22665846299969417, 0.22689620200071658, 0.2267596629990294, 0.2265131450003537], "argmax_stable": false, "spread": {"darwin": 0.0027641616738567527, "einstein": 0.003126653760722969, "hypatia": 0.525409872226754, "ibn_sina": 0.5200232276559871, "sagan": 0.007587552650391804, "socrates": 0.008738026899802496}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.52, "ibn_sina": 0.52, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["ibn_sina", "hypatia", "ibn_sina", "hypatia", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.00723911479158295, "einstein": 0.004621542056406089, "hypatia": 0.1093134469094677, "ibn_sina": 0.8632335117805118, "sagan": 0.007406085461505925, "socrates": 0.008186299000525537}, {"ibn_sina": 0.34321028412452464, "hypatia": 0.6347233191362216, "sagan": 0.00839270414431804, "socrates": 0.003555947391847638, "darwin": 0.0045827331994566425, "einstein": 0.005535012003631481}, {"hypatia": 0.43768400867451257, "darwin": 0.005215214344017496, "sagan": 0.008467405209262634, "ibn_sina": 0.5374190768709828, "einstein": 0.005512546601969112, "socrates": 0.005701748299255441}, {"ibn_sina": 0.4222597458445473, "einstein": 0.004026876093971447, "darwin": 0.005051159796578606, "sagan": 0.014569071716818949, "hypatia": 0.5417991722564337, "socrates": 0.012293974291650134}, {"einstein": 0.007153529854694416, "sagan": 0.014993638111897729, "socrates": 0.0043978544726080045, "ibn_sina": 0.512924022160242, "darwin": 0.005615663697259484, "hypatia": 0.45491529170329836}, {"socrates": 0.006505891892387424, "ibn_sina": 0.8595884126856126, "hypatia": 0.10961254470900345, "einstein": 0.004812330533425867, "sagan": 0.012133925306257254, "darwin": 0.007346894873313395}], "prompt_tokens": 637, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.00723911479158295, "einstein": 0.004621542056406089, "hypatia": 0.1093134469094677, "ibn_sina": 0.8632335117805118, "sagan": 0.007406085461505925, "socrates": 0.008186299000525537}, "confidence": 0.7087805854759335}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.34321028412452464, "hypatia": 0.6347233191362216, "sagan": 0.00839270414431804, "socrates": 0.003555947391847638, "darwin": 0.0045827331994566425, "einstein": 0.005535012003631481}, "confidence": 0.42496205323636504}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.43768400867451257, "darwin": 0.005215214344017496, "sagan": 0.008467405209262634, "ibn_sina": 0.5374190768709828, "einstein": 0.005512546601969112, "socrates": 0.005701748299255441}, "confidence": 0.3766580189795067}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.4222597458445473, "einstein": 0.004026876093971447, "darwin": 0.005051159796578606, "sagan": 0.014569071716818949, "hypatia": 0.5417991722564337, "socrates": 0.012293974291650134}, "confidence": 0.36670571874327845}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.007153529854694416, "sagan": 0.014993638111897729, "socrates": 0.0043978544726080045, "ibn_sina": 0.512924022160242, "darwin": 0.005615663697259484, "hypatia": 0.45491529170329836}, "confidence": 0.36443900093256765}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.006505891892387424, "ibn_sina": 0.8595884126856126, "hypatia": 0.10961254470900345, "einstein": 0.004812330533425867, "sagan": 0.012133925306257254, "darwin": 0.007346894873313395}, "confidence": 0.70140861201166}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:37Z"}
{"id": "c-ibn_sina-01-main-03", "uid": "c063", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "prompt_sha256_by_order": ["9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf"], "choice": "ibn_sina", "probs": {"darwin": 0.015315855939622456, "einstein": 0.019799413034863083, "hypatia": 0.10732533219814058, "ibn_sina": 0.8134214788256653, "sagan": 0.016117204488647938, "socrates": 0.028020715513060734}, "confidence": 0.6098136953493455, "seconds": 0.23092457866672098, "seconds_all_orders": 1.3855474720003258, "seconds_by_order": [0.23091717900024378, 0.23102130799998122, 0.230618351000885, 0.2311811960007617, 0.23088136899968958, 0.2309280689987645], "argmax_stable": true, "spread": {"darwin": 0.011524450762287395, "einstein": 0.015739932372470063, "hypatia": 0.22898759385283093, "ibn_sina": 0.22663991523361848, "sagan": 0.01097303526676249, "socrates": 0.024377562244236026}, "spread_2dp": {"darwin": 0.01, "einstein": 0.009999999999999998, "hypatia": 0.23000000000000004, "ibn_sina": 0.22000000000000008, "sagan": 0.009999999999999998, "socrates": 0.02}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.015315855939622456, "einstein": 0.019799413034863083, "hypatia": 0.10732533219814058, "ibn_sina": 0.8134214788256653, "sagan": 0.016117204488647938, "socrates": 0.028020715513060734}, {"ibn_sina": 0.5867815635920468, "hypatia": 0.3363129260509715, "sagan": 0.021399425738600188, "socrates": 0.01896180141381691, "darwin": 0.011454902884469757, "einstein": 0.02508938032009486}, {"hypatia": 0.15159576829944094, "darwin": 0.017300431361915128, "sagan": 0.02081411025689242, "ibn_sina": 0.7612649259009481, "einstein": 0.022016998776157433, "socrates": 0.027007765404645966}, {"ibn_sina": 0.6864305208037901, "einstein": 0.01733878868591342, "darwin": 0.0182298335586017, "sagan": 0.0186181805008465, "hypatia": 0.21604331279279537, "socrates": 0.04333936365805294}, {"einstein": 0.033078721058383484, "sagan": 0.024067308182788124, "socrates": 0.019859494633510902, "ibn_sina": 0.7877399246237613, "darwin": 0.015048390805374135, "hypatia": 0.12020616069618215}, {"socrates": 0.03134423943758922, "ibn_sina": 0.6503764124784417, "hypatia": 0.24694980371262332, "einstein": 0.021259950969178112, "sagan": 0.027090239755410427, "darwin": 0.022979353646757152}], "prompt_tokens": 683, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.015315855939622456, "einstein": 0.019799413034863083, "hypatia": 0.10732533219814058, "ibn_sina": 0.8134214788256653, "sagan": 0.016117204488647938, "socrates": 0.028020715513060734}, "confidence": 0.6098136953493455}}, "usage": {"input_tokens": 683, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.5867815635920468, "hypatia": 0.3363129260509715, "sagan": 0.021399425738600188, "socrates": 0.01896180141381691, "darwin": 0.011454902884469757, "einstein": 0.02508938032009486}, "confidence": 0.3507965976882777}}, "usage": {"input_tokens": 683, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.15159576829944094, "darwin": 0.017300431361915128, "sagan": 0.02081411025689242, "ibn_sina": 0.7612649259009481, "einstein": 0.022016998776157433, "socrates": 0.027007765404645966}, "confidence": 0.5253427169339557}}, "usage": {"input_tokens": 683, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.6864305208037901, "einstein": 0.01733878868591342, "darwin": 0.0182298335586017, "sagan": 0.0186181805008465, "hypatia": 0.21604331279279537, "socrates": 0.04333936365805294}, "confidence": 0.42486336448667705}}, "usage": {"input_tokens": 683, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.033078721058383484, "sagan": 0.024067308182788124, "socrates": 0.019859494633510902, "ibn_sina": 0.7877399246237613, "darwin": 0.015048390805374135, "hypatia": 0.12020616069618215}, "confidence": 0.5647336007422226}}, "usage": {"input_tokens": 683, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.03134423943758922, "ibn_sina": 0.6503764124784417, "hypatia": 0.24694980371262332, "einstein": 0.021259950969178112, "sagan": 0.027090239755410427, "darwin": 0.022979353646757152}, "confidence": 0.38400407452863095}}, "usage": {"input_tokens": 683, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:38Z"}
{"id": "c-ibn_sina-01-main-03", "uid": "c064", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "prompt_sha256_by_order": ["eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032"], "choice": "ibn_sina", "probs": {"darwin": 0.004204020472574422, "einstein": 0.003542496441763298, "hypatia": 0.2172397694607507, "ibn_sina": 0.6093333801457493, "sagan": 0.0037507078288684208, "socrates": 0.1619296256502938}, "confidence": 0.33369493616420787, "seconds": 0.2276413350000439, "seconds_all_orders": 1.3658480100002635, "seconds_by_order": [0.22760985500099196, 0.22800529199957964, 0.22733313699973223, 0.22754143599922827, 0.227571765999528, 0.22778652400120336], "argmax_stable": false, "spread": {"darwin": 0.0022821202835483128, "einstein": 0.0018148785469654614, "hypatia": 0.40011182585852495, "ibn_sina": 0.2929424049122024, "sagan": 0.0020667890177614827, "socrates": 0.13528465548347757}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.4, "ibn_sina": 0.29, "sagan": 0.01, "socrates": 0.13}, "choices_by_order": ["ibn_sina", "hypatia", "hypatia", "hypatia", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.004204020472574422, "einstein": 0.003542496441763298, "hypatia": 0.2172397694607507, "ibn_sina": 0.6093333801457493, "sagan": 0.0037507078288684208, "socrates": 0.1619296256502938}, {"ibn_sina": 0.3426640992323368, "hypatia": 0.6173515953192756, "sagan": 0.0052315390935710245, "socrates": 0.026644970166816252, "darwin": 0.0034559224359040274, "einstein": 0.004651873752096296}, {"hypatia": 0.48530414491398943, "darwin": 0.0032772167592935034, "sagan": 0.003164750075809542, "ibn_sina": 0.3584268177892935, "einstein": 0.004261518516911061, "socrates": 0.14556555194470303}, {"ibn_sina": 0.3163909752335469, "einstein": 0.004300576133588862, "darwin": 0.0034338091086340405, "sagan": 0.004135639107356903, "hypatia": 0.6085436165219444, "socrates": 0.06319538389492879}, {"einstein": 0.005357374988728759, "sagan": 0.004270225359916675, "socrates": 0.033807181475777416, "ibn_sina": 0.5168135118484876, "darwin": 0.005559337042841816, "hypatia": 0.4341923692842478}, {"socrates": 0.06875646235406535, "ibn_sina": 0.38937288379622187, "hypatia": 0.5303432529759249, "einstein": 0.004365696518737234, "sagan": 0.003566957009934121, "darwin": 0.0035947473451164785}], "prompt_tokens": 647, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.004204020472574422, "einstein": 0.003542496441763298, "hypatia": 0.2172397694607507, "ibn_sina": 0.6093333801457493, "sagan": 0.0037507078288684208, "socrates": 0.1619296256502938}, "confidence": 0.33369493616420787}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.3426640992323368, "hypatia": 0.6173515953192756, "sagan": 0.0052315390935710245, "socrates": 0.026644970166816252, "darwin": 0.0034559224359040274, "einstein": 0.004651873752096296}, "confidence": 0.39917510069407597}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.48530414491398943, "darwin": 0.0032772167592935034, "sagan": 0.003164750075809542, "ibn_sina": 0.3584268177892935, "einstein": 0.004261518516911061, "socrates": 0.14556555194470303}, "confidence": 0.26226177163242825}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.3163909752335469, "einstein": 0.004300576133588862, "darwin": 0.0034338091086340405, "sagan": 0.004135639107356903, "hypatia": 0.6085436165219444, "socrates": 0.06319538389492879}, "confidence": 0.3693635541704456}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.005357374988728759, "sagan": 0.004270225359916675, "socrates": 0.033807181475777416, "ibn_sina": 0.5168135118484876, "darwin": 0.005559337042841816, "hypatia": 0.4341923692842478}, "confidence": 0.34820798513550616}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"socrates": 0.06875646235406535, "ibn_sina": 0.38937288379622187, "hypatia": 0.5303432529759249, "einstein": 0.004365696518737234, "sagan": 0.003566957009934121, "darwin": 0.0035947473451164785}, "confidence": 0.32517683731023717}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:39Z"}
{"id": "c-ibn_sina-01-main-03", "uid": "c065", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "prompt_sha256_by_order": ["c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086"], "choice": "ibn_sina", "probs": {"darwin": 0.01147439094472828, "einstein": 0.03716346946955847, "hypatia": 0.010710680339829293, "ibn_sina": 0.8832817119168507, "sagan": 0.038581776381183416, "socrates": 0.018787970947849918}, "confidence": 0.740386753254084, "seconds": 0.25386347116697533, "seconds_all_orders": 1.5231808270018519, "seconds_by_order": [0.2538729190000595, 0.2538612899988948, 0.25347025300106907, 0.254627074000382, 0.2535565520011005, 0.25379273900034605], "argmax_stable": true, "spread": {"darwin": 0.004008674610135614, "einstein": 0.09607391240259053, "hypatia": 0.006788669163999131, "ibn_sina": 0.08108310672474062, "sagan": 0.048223911414770546, "socrates": 0.014625208854906327}, "spread_2dp": {"darwin": 0.0, "einstein": 0.09999999999999999, "hypatia": 0.01, "ibn_sina": 0.08000000000000007, "sagan": 0.049999999999999996, "socrates": 0.01}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.01147439094472828, "einstein": 0.03716346946955847, "hypatia": 0.010710680339829293, "ibn_sina": 0.8832817119168507, "sagan": 0.038581776381183416, "socrates": 0.018787970947849918}, {"ibn_sina": 0.8594945075294192, "hypatia": 0.013708431740363129, "sagan": 0.0526847981694682, "socrates": 0.006130410224066743, "darwin": 0.010752420053359378, "einstein": 0.057229432283323306}, {"hypatia": 0.016034919230076224, "darwin": 0.012761373380844912, "sagan": 0.013788749713676586, "ibn_sina": 0.8946850781171847, "einstein": 0.051599016052063636, "socrates": 0.011130863506154005}, {"ibn_sina": 0.8494641573803389, "einstein": 0.047747223197381526, "darwin": 0.00921137526864126, "sagan": 0.06201266112844713, "hypatia": 0.010808963946218107, "socrates": 0.02075561907897307}, {"einstein": 0.11877512849458109, "sagan": 0.0197697405396985, "socrates": 0.0076928627153871314, "ibn_sina": 0.8313933345289437, "darwin": 0.010796081642462697, "hypatia": 0.011572852078927005}, {"socrates": 0.01189438816794949, "ibn_sina": 0.9124764412536843, "hypatia": 0.009246250066077093, "einstein": 0.022701216091990557, "sagan": 0.030461654541521756, "darwin": 0.013220049878776874}], "prompt_tokens": 693, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.01147439094472828, "einstein": 0.03716346946955847, "hypatia": 0.010710680339829293, "ibn_sina": 0.8832817119168507, "sagan": 0.038581776381183416, "socrates": 0.018787970947849918}, "confidence": 0.740386753254084}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.8594945075294192, "hypatia": 0.013708431740363129, "sagan": 0.0526847981694682, "socrates": 0.006130410224066743, "darwin": 0.010752420053359378, "einstein": 0.057229432283323306}, "confidence": 0.6941473863018007}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.016034919230076224, "darwin": 0.012761373380844912, "sagan": 0.013788749713676586, "ibn_sina": 0.8946850781171847, "einstein": 0.051599016052063636, "socrates": 0.011130863506154005}, "confidence": 0.7646294133873202}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.8494641573803389, "einstein": 0.047747223197381526, "darwin": 0.00921137526864126, "sagan": 0.06201266112844713, "hypatia": 0.010808963946218107, "socrates": 0.02075561907897307}, "confidence": 0.6740166411957557}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.11877512849458109, "sagan": 0.0197697405396985, "socrates": 0.0076928627153871314, "ibn_sina": 0.8313933345289437, "darwin": 0.010796081642462697, "hypatia": 0.011572852078927005}, "confidence": 0.6472275002918758}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.01189438816794949, "ibn_sina": 0.9124764412536843, "hypatia": 0.009246250066077093, "einstein": 0.022701216091990557, "sagan": 0.030461654541521756, "darwin": 0.013220049878776874}, "confidence": 0.8013499033377067}}, "usage": {"input_tokens": 693, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:41Z"}
{"id": "c-ibn_sina-01-starter-01", "uid": "c066", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "prompt_sha256_by_order": ["43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423"], "choice": "hypatia", "probs": {"darwin": 0.005475786557178668, "einstein": 0.004555127197569244, "hypatia": 0.4921662423572172, "ibn_sina": 0.46589698252602774, "sagan": 0.005265211829316989, "socrates": 0.026640649532690137}, "confidence": 0.3520909462338614, "seconds": 0.19303760183326327, "seconds_all_orders": 1.1582256109995797, "seconds_by_order": [0.192797932000758, 0.19311990999995032, 0.19289216099969053, 0.1932632680000097, 0.19319857899972703, 0.19295376099944406], "argmax_stable": false, "spread": {"darwin": 0.0029415380997511323, "einstein": 0.009034287743702291, "hypatia": 0.15965852882056664, "ibn_sina": 0.14283008600658592, "sagan": 0.003907535652010437, "socrates": 0.023597838171138565}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.16000000000000003, "ibn_sina": 0.15000000000000002, "sagan": 0.01, "socrates": 0.019999999999999997}, "choices_by_order": ["hypatia", "hypatia", "hypatia", "hypatia", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.005475786557178668, "einstein": 0.004555127197569244, "hypatia": 0.4921662423572172, "ibn_sina": 0.46589698252602774, "sagan": 0.005265211829316989, "socrates": 0.026640649532690137}, {"ibn_sina": 0.35229511282107745, "hypatia": 0.6255925905167026, "sagan": 0.0062250215362495, "socrates": 0.0070834921682404435, "darwin": 0.005081933379517291, "einstein": 0.0037218495782125466}, {"hypatia": 0.4807033230518239, "darwin": 0.00525980445611483, "sagan": 0.006778768393598599, "ibn_sina": 0.47138319982412324, "einstein": 0.005193573934960385, "socrates": 0.030681330339379007}, {"ibn_sina": 0.365199053039863, "einstein": 0.0050714816978588, "darwin": 0.005292597952764787, "sagan": 0.005586886174578833, "hypatia": 0.5902621143089309, "socrates": 0.02858786682600363}, {"einstein": 0.012756137321914838, "sagan": 0.007394291337155279, "socrates": 0.011223557179883612, "ibn_sina": 0.49512519882766337, "darwin": 0.00756675363724705, "hypatia": 0.46593406169613594}, {"socrates": 0.023111303123223646, "ibn_sina": 0.4318120594047842, "hypatia": 0.531759256537157, "einstein": 0.005205409712194348, "sagan": 0.0034867556851448417, "darwin": 0.004625215537495918}], "prompt_tokens": 493, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"darwin": 0.005475786557178668, "einstein": 0.004555127197569244, "hypatia": 0.4921662423572172, "ibn_sina": 0.46589698252602774, "sagan": 0.005265211829316989, "socrates": 0.026640649532690137}, "confidence": 0.3520909462338614}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.35229511282107745, "hypatia": 0.6255925905167026, "sagan": 0.0062250215362495, "socrates": 0.0070834921682404435, "darwin": 0.005081933379517291, "einstein": 0.0037218495782125466}, "confidence": 0.4187278489511145}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.4807033230518239, "darwin": 0.00525980445611483, "sagan": 0.006778768393598599, "ibn_sina": 0.47138319982412324, "einstein": 0.005193573934960385, "socrates": 0.030681330339379007}, "confidence": 0.3451836884259497}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.365199053039863, "einstein": 0.0050714816978588, "darwin": 0.005292597952764787, "sagan": 0.005586886174578833, "hypatia": 0.5902621143089309, "socrates": 0.02858786682600363}, "confidence": 0.379218307451383}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.012756137321914838, "sagan": 0.007394291337155279, "socrates": 0.011223557179883612, "ibn_sina": 0.49512519882766337, "darwin": 0.00756675363724705, "hypatia": 0.46593406169613594}, "confidence": 0.3551729571315076}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"socrates": 0.023111303123223646, "ibn_sina": 0.4318120594047842, "hypatia": 0.531759256537157, "einstein": 0.005205409712194348, "sagan": 0.0034867556851448417, "darwin": 0.004625215537495918}, "confidence": 0.36378920831998285}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:42Z"}
{"id": "c-ibn_sina-01-starter-02", "uid": "c067", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "prompt_sha256_by_order": ["131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b"], "choice": "einstein", "probs": {"darwin": 0.08176851517962441, "einstein": 0.40098779663098505, "hypatia": 0.019670138782556267, "ibn_sina": 0.05893396413491145, "sagan": 0.18788027147386163, "socrates": 0.25075931379806127}, "confidence": 0.1234199913707191, "seconds": 0.18912023283337476, "seconds_all_orders": 1.1347213970002485, "seconds_by_order": [0.18898766400161549, 0.18938345999958983, 0.18908692299919494, 0.18946047999997973, 0.18889349500022945, 0.1889093749996391], "argmax_stable": false, "spread": {"darwin": 0.04861512782641271, "einstein": 0.27456771520362483, "hypatia": 0.008780167191695094, "ibn_sina": 0.03288228468473911, "sagan": 0.17289633435259313, "socrates": 0.1652826872114185}, "spread_2dp": {"darwin": 0.039999999999999994, "einstein": 0.2700000000000001, "hypatia": 0.01, "ibn_sina": 0.030000000000000006, "sagan": 0.16999999999999998, "socrates": 0.16}, "choices_by_order": ["einstein", "einstein", "einstein", "sagan", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.08176851517962441, "einstein": 0.40098779663098505, "hypatia": 0.019670138782556267, "ibn_sina": 0.05893396413491145, "sagan": 0.18788027147386163, "socrates": 0.25075931379806127}, {"ibn_sina": 0.061775641081892305, "hypatia": 0.018743019085028168, "sagan": 0.3388234142282637, "socrates": 0.08547662658664276, "darwin": 0.06912029112525887, "einstein": 0.42606100789291423}, {"hypatia": 0.014479718908441757, "darwin": 0.07371737422946191, "sagan": 0.29474673479489544, "ibn_sina": 0.04158984667941734, "einstein": 0.37145159248988957, "socrates": 0.20401473289789393}, {"ibn_sina": 0.051103715296359105, "einstein": 0.28818199650078274, "darwin": 0.06633402839102587, "sagan": 0.36077660582645477, "hypatia": 0.022876611452599047, "socrates": 0.21072704253277844}, {"einstein": 0.5627497117044076, "sagan": 0.19528672822517604, "socrates": 0.0874417066609743, "ibn_sina": 0.05908915734494617, "darwin": 0.08011404894937356, "hypatia": 0.015318647115122291}, {"socrates": 0.1677027483970246, "ibn_sina": 0.07447213136415645, "hypatia": 0.02325988610013685, "einstein": 0.35480091190440205, "sagan": 0.26481516601684146, "darwin": 0.11494915621743858}], "prompt_tokens": 529, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.08176851517962441, "einstein": 0.40098779663098505, "hypatia": 0.019670138782556267, "ibn_sina": 0.05893396413491145, "sagan": 0.18788027147386163, "socrates": 0.25075931379806127}, "confidence": 0.1234199913707191}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.061775641081892305, "hypatia": 0.018743019085028168, "sagan": 0.3388234142282637, "socrates": 0.08547662658664276, "darwin": 0.06912029112525887, "einstein": 0.42606100789291423}, "confidence": 0.17509682489127631}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.014479718908441757, "darwin": 0.07371737422946191, "sagan": 0.29474673479489544, "ibn_sina": 0.04158984667941734, "einstein": 0.37145159248988957, "socrates": 0.20401473289789393}, "confidence": 0.12861707601363745}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.051103715296359105, "einstein": 0.28818199650078274, "darwin": 0.06633402839102587, "sagan": 0.36077660582645477, "hypatia": 0.022876611452599047, "socrates": 0.21072704253277844}, "confidence": 0.11817956951760591}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.5627497117044076, "sagan": 0.19528672822517604, "socrates": 0.0874417066609743, "ibn_sina": 0.05908915734494617, "darwin": 0.08011404894937356, "hypatia": 0.015318647115122291}, "confidence": 0.24713757593476907}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.1677027483970246, "ibn_sina": 0.07447213136415645, "hypatia": 0.02325988610013685, "einstein": 0.35480091190440205, "sagan": 0.26481516601684146, "darwin": 0.11494915621743858}, "confidence": 0.09212248027246109}}, "usage": {"input_tokens": 529, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:43Z"}
{"id": "c-ibn_sina-01-starter-03", "uid": "c068", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "prompt_sha256_by_order": ["34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf"], "choice": "ibn_sina", "probs": {"darwin": 0.008621283009511553, "einstein": 0.00940852122162981, "hypatia": 0.10316221341262258, "ibn_sina": 0.8564484481688643, "sagan": 0.010630810848989846, "socrates": 0.01172872333838189}, "confidence": 0.6934717726357726, "seconds": 0.22718590249981693, "seconds_all_orders": 1.3631154149989015, "seconds_by_order": [0.22743893700135231, 0.2274305369992362, 0.22656183499930194, 0.22721509900111414, 0.2273463379988243, 0.22712266899907263], "argmax_stable": true, "spread": {"darwin": 0.006797991696511211, "einstein": 0.0057256802850622675, "hypatia": 0.1337342646356262, "ibn_sina": 0.15049803098964887, "sagan": 0.004449291730878727, "socrates": 0.008121939272416855}, "spread_2dp": {"darwin": 0.0, "einstein": 0.0, "hypatia": 0.13, "ibn_sina": 0.15000000000000002, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.008621283009511553, "einstein": 0.00940852122162981, "hypatia": 0.10316221341262258, "ibn_sina": 0.8564484481688643, "sagan": 0.010630810848989846, "socrates": 0.01172872333838189}, {"ibn_sina": 0.8766495197240665, "hypatia": 0.09550003617002985, "sagan": 0.008110680915606332, "socrates": 0.006347091537840747, "darwin": 0.005549896439059574, "einstein": 0.007842775213397052}, {"hypatia": 0.09958043366396122, "darwin": 0.010498922652005789, "sagan": 0.01193281697973899, "ibn_sina": 0.8563292755383398, "einstein": 0.01230580615029413, "socrates": 0.009352745015660103}, {"ibn_sina": 0.7261514887344176, "einstein": 0.01104038486181005, "darwin": 0.00925262479665242, "sagan": 0.009852169991206354, "hypatia": 0.22923430080565604, "socrates": 0.014469030810257601}, {"einstein": 0.013161859027815744, "sagan": 0.011183943272884207, "socrates": 0.01040071811511369, "ibn_sina": 0.841903143281679, "darwin": 0.012347888135570785, "hypatia": 0.11100244816693657}, {"socrates": 0.009141539841082003, "ibn_sina": 0.869525769780534, "hypatia": 0.1005792216244006, "einstein": 0.007436178742753477, "sagan": 0.0074835252488602635, "darwin": 0.005833764762369718}], "prompt_tokens": 645, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.008621283009511553, "einstein": 0.00940852122162981, "hypatia": 0.10316221341262258, "ibn_sina": 0.8564484481688643, "sagan": 0.010630810848989846, "socrates": 0.01172872333838189}, "confidence": 0.6934717726357726}}, "usage": {"input_tokens": 645, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.8766495197240665, "hypatia": 0.09550003617002985, "sagan": 0.008110680915606332, "socrates": 0.006347091537840747, "darwin": 0.005549896439059574, "einstein": 0.007842775213397052}, "confidence": 0.7333996198364201}}, "usage": {"input_tokens": 645, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.09958043366396122, "darwin": 0.010498922652005789, "sagan": 0.01193281697973899, "ibn_sina": 0.8563292755383398, "einstein": 0.01230580615029413, "socrates": 0.009352745015660103}, "confidence": 0.6924491405379782}}, "usage": {"input_tokens": 645, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.7261514887344176, "einstein": 0.01104038486181005, "darwin": 0.00925262479665242, "sagan": 0.009852169991206354, "hypatia": 0.22923430080565604, "socrates": 0.014469030810257601}, "confidence": 0.49642992223206067}}, "usage": {"input_tokens": 645, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.013161859027815744, "sagan": 0.011183943272884207, "socrates": 0.01040071811511369, "ibn_sina": 0.841903143281679, "darwin": 0.012347888135570785, "hypatia": 0.11100244816693657}, "confidence": 0.6660176878786512}}, "usage": {"input_tokens": 645, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.009141539841082003, "ibn_sina": 0.869525769780534, "hypatia": 0.1005792216244006, "einstein": 0.007436178742753477, "sagan": 0.0074835252488602635, "darwin": 0.005833764762369718}, "confidence": 0.7197041735216965}}, "usage": {"input_tokens": 645, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:45Z"}
{"id": "c-ibn_sina-01-starter-03", "uid": "c069", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "prompt_sha256_by_order": ["278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442"], "choice": "ibn_sina", "probs": {"darwin": 0.04598919894788146, "einstein": 0.024829798263562297, "hypatia": 0.028423496738928303, "ibn_sina": 0.8683753438065009, "sagan": 0.01866443412404125, "socrates": 0.01371772811908577}, "confidence": 0.7097820344386282, "seconds": 0.22457495549982318, "seconds_all_orders": 1.347449732998939, "seconds_by_order": [0.22439225199923385, 0.2247291090006911, 0.22414046400081133, 0.2248719279996294, 0.22494831799849635, 0.22436766200007696], "argmax_stable": true, "spread": {"darwin": 0.017654224627150994, "einstein": 0.015024468921487493, "hypatia": 0.04211592167479474, "ibn_sina": 0.04368290807333686, "sagan": 0.0077806199956778, "socrates": 0.010324207624139797}, "spread_2dp": {"darwin": 0.009999999999999995, "einstein": 0.019999999999999997, "hypatia": 0.04000000000000001, "ibn_sina": 0.040000000000000036, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.04598919894788146, "einstein": 0.024829798263562297, "hypatia": 0.028423496738928303, "ibn_sina": 0.8683753438065009, "sagan": 0.01866443412404125, "socrates": 0.01371772811908577}, {"ibn_sina": 0.8321641704838236, "hypatia": 0.0460700876291785, "sagan": 0.020184229737243844, "socrates": 0.008418719319372383, "darwin": 0.06364342357503246, "einstein": 0.029519369255349082}, {"hypatia": 0.06804326247183343, "darwin": 0.06051947797917983, "sagan": 0.016326930468319383, "ibn_sina": 0.8279332269503284, "einstein": 0.0176977306066997, "socrates": 0.009479371523639181}, {"ibn_sina": 0.8371724280376307, "einstein": 0.0240799845216339, "darwin": 0.047904739776101185, "sagan": 0.015116594292362453, "hypatia": 0.05846182606230495, "socrates": 0.01726442730996692}, {"einstein": 0.022615226428298857, "sagan": 0.013761416110852222, "socrates": 0.006940219685827123, "ibn_sina": 0.8385896872521972, "darwin": 0.055919434888598796, "hypatia": 0.062174015634225796}, {"socrates": 0.014174112536281234, "ibn_sina": 0.8716161350236653, "hypatia": 0.025927340797038685, "einstein": 0.01449490033386159, "sagan": 0.02154203610653002, "darwin": 0.05224547520262323}], "prompt_tokens": 617, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.04598919894788146, "einstein": 0.024829798263562297, "hypatia": 0.028423496738928303, "ibn_sina": 0.8683753438065009, "sagan": 0.01866443412404125, "socrates": 0.01371772811908577}, "confidence": 0.7097820344386282}}, "usage": {"input_tokens": 617, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.8321641704838236, "hypatia": 0.0460700876291785, "sagan": 0.020184229737243844, "socrates": 0.008418719319372383, "darwin": 0.06364342357503246, "einstein": 0.029519369255349082}, "confidence": 0.6400237793222072}}, "usage": {"input_tokens": 617, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.06804326247183343, "darwin": 0.06051947797917983, "sagan": 0.016326930468319383, "ibn_sina": 0.8279332269503284, "einstein": 0.0176977306066997, "socrates": 0.009479371523639181}, "confidence": 0.6333226694592342}}, "usage": {"input_tokens": 617, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.8371724280376307, "einstein": 0.0240799845216339, "darwin": 0.047904739776101185, "sagan": 0.015116594292362453, "hypatia": 0.05846182606230495, "socrates": 0.01726442730996692}, "confidence": 0.6492120891926565}}, "usage": {"input_tokens": 617, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.022615226428298857, "sagan": 0.013761416110852222, "socrates": 0.006940219685827123, "ibn_sina": 0.8385896872521972, "darwin": 0.055919434888598796, "hypatia": 0.062174015634225796}, "confidence": 0.65316909600778}}, "usage": {"input_tokens": 617, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.014174112536281234, "ibn_sina": 0.8716161350236653, "hypatia": 0.025927340797038685, "einstein": 0.01449490033386159, "sagan": 0.02154203610653002, "darwin": 0.05224547520262323}, "confidence": 0.7167898845220491}}, "usage": {"input_tokens": 617, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:46Z"}
{"id": "c-ibn_sina-01-starter-03", "uid": "c070", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "prompt_sha256_by_order": ["1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728"], "choice": "einstein", "probs": {"darwin": 0.005312919784762814, "einstein": 0.7262698038442612, "hypatia": 0.006378004148013259, "ibn_sina": 0.010553358328613083, "sagan": 0.2466655085991169, "socrates": 0.0048204052952328545}, "confidence": 0.5062182602094598, "seconds": 0.22746445316715835, "seconds_all_orders": 1.3647867190029501, "seconds_by_order": [0.22686067199902027, 0.22787787299967022, 0.22788898300132132, 0.22691375100112054, 0.22783996300131548, 0.2274054770005023], "argmax_stable": true, "spread": {"darwin": 0.0014628767745965207, "einstein": 0.18549294467053234, "hypatia": 0.002123998681347883, "ibn_sina": 0.0076888920482871765, "sagan": 0.19279877680206553, "socrates": 0.001268988645437794}, "spread_2dp": {"darwin": 0.01, "einstein": 0.18999999999999995, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.19, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.005312919784762814, "einstein": 0.7262698038442612, "hypatia": 0.006378004148013259, "ibn_sina": 0.010553358328613083, "sagan": 0.2466655085991169, "socrates": 0.0048204052952328545}, {"ibn_sina": 0.013303389662858286, "hypatia": 0.005995765376645958, "sagan": 0.327185753938333, "socrates": 0.005281040325840038, "darwin": 0.0052328210290884004, "einstein": 0.6430012296672343}, {"hypatia": 0.005182188176685889, "darwin": 0.004655357440648957, "sagan": 0.35494256969819926, "ibn_sina": 0.0064606209307506955, "einstein": 0.623525783709334, "socrates": 0.0052334800443811814}, {"ibn_sina": 0.007476427151270215, "einstein": 0.6654822422083742, "darwin": 0.005264626881419057, "sagan": 0.3113880961922706, "hypatia": 0.006328973483599494, "socrates": 0.004059634083066525}, {"einstein": 0.5407768591737289, "sagan": 0.43946428540118243, "socrates": 0.005328622728504319, "ibn_sina": 0.00561449761457111, "darwin": 0.004561729615347881, "hypatia": 0.004254005466665376}, {"socrates": 0.0045112488584598785, "ibn_sina": 0.010349743043247577, "hypatia": 0.005230596768325074, "einstein": 0.6330151521675377, "sagan": 0.3408686527724854, "darwin": 0.0060246063899444015}], "prompt_tokens": 647, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.005312919784762814, "einstein": 0.7262698038442612, "hypatia": 0.006378004148013259, "ibn_sina": 0.010553358328613083, "sagan": 0.2466655085991169, "socrates": 0.0048204052952328545}, "confidence": 0.5062182602094598}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.013303389662858286, "hypatia": 0.005995765376645958, "sagan": 0.327185753938333, "socrates": 0.005281040325840038, "darwin": 0.0052328210290884004, "einstein": 0.6430012296672343}, "confidence": 0.42492316013870796}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.005182188176685889, "darwin": 0.004655357440648957, "sagan": 0.35494256969819926, "ibn_sina": 0.0064606209307506955, "einstein": 0.623525783709334, "socrates": 0.0052334800443811814}, "confidence": 0.41786354451722163}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.007476427151270215, "einstein": 0.6654822422083742, "darwin": 0.005264626881419057, "sagan": 0.3113880961922706, "hypatia": 0.006328973483599494, "socrates": 0.004059634083066525}, "confidence": 0.44796317312599027}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.5407768591737289, "sagan": 0.43946428540118243, "socrates": 0.005328622728504319, "ibn_sina": 0.00561449761457111, "darwin": 0.004561729615347881, "hypatia": 0.004254005466665376}, "confidence": 0.38280075076505876}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.0045112488584598785, "ibn_sina": 0.010349743043247577, "hypatia": 0.005230596768325074, "einstein": 0.6330151521675377, "sagan": 0.3408686527724854, "darwin": 0.0060246063899444015}, "confidence": 0.42050889386638096}}, "usage": {"input_tokens": 647, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:47Z"}
{"id": "c-ibn_sina-01-starter-03", "uid": "c071", "task": "kperm", "label": "ibn_sina", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "prompt_sha256_by_order": ["ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c"], "choice": "ibn_sina", "probs": {"darwin": 0.008764918724028629, "einstein": 0.009241465843456894, "hypatia": 0.43444797951693837, "ibn_sina": 0.4827720070193957, "sagan": 0.028114528938795784, "socrates": 0.03665909995738455}, "confidence": 0.30893248300743525, "seconds": 0.22301262349992612, "seconds_all_orders": 1.3380757409995567, "seconds_by_order": [0.22358746899953985, 0.2226204559992766, 0.22284182499970484, 0.2226592460010579, 0.22360681899954216, 0.22275992600043537], "argmax_stable": false, "spread": {"darwin": 0.0036604818030788605, "einstein": 0.00837700262465159, "hypatia": 0.3186717286938058, "ibn_sina": 0.36618587853701784, "sagan": 0.09546648908949396, "socrates": 0.026515308445881285}, "spread_2dp": {"darwin": 0.0, "einstein": 0.01, "hypatia": 0.32, "ibn_sina": 0.36, "sagan": 0.09, "socrates": 0.03}, "choices_by_order": ["ibn_sina", "hypatia", "hypatia", "hypatia", "hypatia", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.008764918724028629, "einstein": 0.009241465843456894, "hypatia": 0.43444797951693837, "ibn_sina": 0.4827720070193957, "sagan": 0.028114528938795784, "socrates": 0.03665909995738455}, {"ibn_sina": 0.11658612848237786, "hypatia": 0.7531197082107441, "sagan": 0.09010650280015876, "socrates": 0.010143791511503264, "darwin": 0.01242540052710749, "einstein": 0.017618468468108484}, {"hypatia": 0.572600602314873, "darwin": 0.010029392032003558, "sagan": 0.06583934255646287, "ibn_sina": 0.30604014959894627, "einstein": 0.014193748746355844, "socrates": 0.031296764751358536}, {"ibn_sina": 0.26698032686323464, "einstein": 0.009465435963420945, "darwin": 0.010173070922415511, "sagan": 0.0659756632511535, "hypatia": 0.6262266148459258, "socrates": 0.021178888153849593}, {"einstein": 0.016628561279346747, "sagan": 0.12358101802828975, "socrates": 0.013685097455227278, "ibn_sina": 0.17738382268635036, "darwin": 0.01109822193717909, "hypatia": 0.6576232786136068}, {"socrates": 0.019441930894382658, "ibn_sina": 0.16736123603352457, "hypatia": 0.7470003557625122, "einstein": 0.009961033351586983, "sagan": 0.046150287348444646, "darwin": 0.010085156609548961}], "prompt_tokens": 599, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.008764918724028629, "einstein": 0.009241465843456894, "hypatia": 0.43444797951693837, "ibn_sina": 0.4827720070193957, "sagan": 0.028114528938795784, "socrates": 0.03665909995738455}, "confidence": 0.30893248300743525}}, "usage": {"input_tokens": 599, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.11658612848237786, "hypatia": 0.7531197082107441, "sagan": 0.09010650280015876, "socrates": 0.010143791511503264, "darwin": 0.01242540052710749, "einstein": 0.017618468468108484}, "confidence": 0.5073621995349645}}, "usage": {"input_tokens": 599, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"hypatia": 0.572600602314873, "darwin": 0.010029392032003558, "sagan": 0.06583934255646287, "ibn_sina": 0.30604014959894627, "einstein": 0.014193748746355844, "socrates": 0.031296764751358536}, "confidence": 0.3125780567896914}}, "usage": {"input_tokens": 599, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"ibn_sina": 0.26698032686323464, "einstein": 0.009465435963420945, "darwin": 0.010173070922415511, "sagan": 0.0659756632511535, "hypatia": 0.6262266148459258, "socrates": 0.021178888153849593}, "confidence": 0.3621192248418642}}, "usage": {"input_tokens": 599, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.016628561279346747, "sagan": 0.12358101802828975, "socrates": 0.013685097455227278, "ibn_sina": 0.17738382268635036, "darwin": 0.01109822193717909, "hypatia": 0.6576232786136068}, "confidence": 0.37575115193798503}}, "usage": {"input_tokens": 599, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"socrates": 0.019441930894382658, "ibn_sina": 0.16736123603352457, "hypatia": 0.7470003557625122, "einstein": 0.009961033351586983, "sagan": 0.046150287348444646, "darwin": 0.010085156609548961}, "confidence": 0.5064737021254067}}, "usage": {"input_tokens": 599, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:49Z"}
{"id": "c-sagan-01-dessert-01", "uid": "c072", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "prompt_sha256_by_order": ["66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f"], "choice": "sagan", "probs": {"darwin": 0.016198408442118624, "einstein": 0.38989628534759885, "hypatia": 0.010203131713092006, "ibn_sina": 0.009741367071606463, "sagan": 0.5508064867255675, "socrates": 0.023154320700016606}, "confidence": 0.3476852899349613, "seconds": 0.25525575499978004, "seconds_all_orders": 1.5315345299986802, "seconds_by_order": [0.25491690099988773, 0.2554201260008995, 0.2562611689991172, 0.25492594099887356, 0.2550358709995635, 0.25497452200033877], "argmax_stable": false, "spread": {"darwin": 0.013293532390255772, "einstein": 0.4294393108681503, "hypatia": 0.007827525728626659, "ibn_sina": 0.00836606895772531, "sagan": 0.42824965124463077, "socrates": 0.021896507971751994}, "spread_2dp": {"darwin": 0.01, "einstein": 0.42999999999999994, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.43000000000000005, "socrates": 0.03}, "choices_by_order": ["sagan", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.016198408442118624, "einstein": 0.38989628534759885, "hypatia": 0.010203131713092006, "ibn_sina": 0.009741367071606463, "sagan": 0.5508064867255675, "socrates": 0.023154320700016606}, {"ibn_sina": 0.009994459319652624, "hypatia": 0.008897826195337823, "sagan": 0.12255683548093677, "socrates": 0.016438726425625154, "darwin": 0.0227765563626984, "einstein": 0.8193355962157491}, {"hypatia": 0.009514220220527655, "darwin": 0.012285825589944028, "sagan": 0.21462094008865118, "ibn_sina": 0.011398203777075979, "einstein": 0.7226466175326482, "socrates": 0.029534192791152998}, {"ibn_sina": 0.01594862791543329, "einstein": 0.549673824237101, "darwin": 0.01576944923586196, "sagan": 0.36686597053285935, "hypatia": 0.016017813208366698, "socrates": 0.035724314870377784}, {"einstein": 0.7265670470339545, "sagan": 0.2343492756575292, "socrates": 0.01382780689862579, "ibn_sina": 0.007582558957707982, "darwin": 0.009483023972442627, "hypatia": 0.008190287479740039}, {"socrates": 0.03086001612509142, "ibn_sina": 0.013699998902334223, "hypatia": 0.009979565115651087, "einstein": 0.5154387376787568, "sagan": 0.4148366094796783, "darwin": 0.015185072698488152}], "prompt_tokens": 709, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.016198408442118624, "einstein": 0.38989628534759885, "hypatia": 0.010203131713092006, "ibn_sina": 0.009741367071606463, "sagan": 0.5508064867255675, "socrates": 0.023154320700016606}, "confidence": 0.3476852899349613}}, "usage": {"input_tokens": 709, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.009994459319652624, "hypatia": 0.008897826195337823, "sagan": 0.12255683548093677, "socrates": 0.016438726425625154, "darwin": 0.0227765563626984, "einstein": 0.8193355962157491}, "confidence": 0.6247588731083834}}, "usage": {"input_tokens": 709, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.009514220220527655, "darwin": 0.012285825589944028, "sagan": 0.21462094008865118, "ibn_sina": 0.011398203777075979, "einstein": 0.7226466175326482, "socrates": 0.029534192791152998}, "confidence": 0.48342871749509675}}, "usage": {"input_tokens": 709, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.01594862791543329, "einstein": 0.549673824237101, "darwin": 0.01576944923586196, "sagan": 0.36686597053285935, "hypatia": 0.016017813208366698, "socrates": 0.035724314870377784}, "confidence": 0.3265213415931755}}, "usage": {"input_tokens": 709, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.7265670470339545, "sagan": 0.2343492756575292, "socrates": 0.01382780689862579, "ibn_sina": 0.007582558957707982, "darwin": 0.009483023972442627, "hypatia": 0.008190287479740039}, "confidence": 0.4998699626001817}}, "usage": {"input_tokens": 709, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.03086001612509142, "ibn_sina": 0.013699998902334223, "hypatia": 0.009979565115651087, "einstein": 0.5154387376787568, "sagan": 0.4148366094796783, "darwin": 0.015185072698488152}, "confidence": 0.3270840562988634}}, "usage": {"input_tokens": 709, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:50Z"}
{"id": "c-sagan-01-dessert-01", "uid": "c073", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "prompt_sha256_by_order": ["985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549"], "choice": "sagan", "probs": {"darwin": 0.04232267771053761, "einstein": 0.025590383371811, "hypatia": 0.007144521263875355, "ibn_sina": 0.010485440165061352, "sagan": 0.8908750666343173, "socrates": 0.023581910854397307}, "confidence": 0.7561858675353944, "seconds": 0.1883539330004472, "seconds_all_orders": 1.1301235980026831, "seconds_by_order": [0.1889361840003403, 0.18789741300133755, 0.1886595470004977, 0.18791758400038816, 0.18847648900009517, 0.18823638100002427], "argmax_stable": true, "spread": {"darwin": 0.03182359412563802, "einstein": 0.009108666119507305, "hypatia": 0.003542596192640182, "ibn_sina": 0.0031286935296968785, "sagan": 0.03928851086080887, "socrates": 0.018725056368357536}, "spread_2dp": {"darwin": 0.03, "einstein": 0.009999999999999998, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.040000000000000036, "socrates": 0.02}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.04232267771053761, "einstein": 0.025590383371811, "hypatia": 0.007144521263875355, "ibn_sina": 0.010485440165061352, "sagan": 0.8908750666343173, "socrates": 0.023581910854397307}, {"ibn_sina": 0.010030933232457169, "hypatia": 0.008423258504315159, "sagan": 0.9230459027590482, "socrates": 0.02646482751432718, "darwin": 0.015553360737548604, "einstein": 0.016481717252303695}, {"hypatia": 0.007279833257480972, "darwin": 0.015912985093585696, "sagan": 0.9046288193968632, "ibn_sina": 0.010990038830519813, "einstein": 0.018881356198795482, "socrates": 0.04230696722275484}, {"ibn_sina": 0.011307818645799855, "einstein": 0.023646815020430846, "darwin": 0.017448949889511577, "sagan": 0.9116520150787981, "hypatia": 0.009738532068183972, "socrates": 0.026205869297275718}, {"einstein": 0.01715740774084351, "sagan": 0.9301635774951261, "socrates": 0.027046784479573635, "ibn_sina": 0.00889261518480956, "darwin": 0.010499083584899593, "hypatia": 0.0062405315147475065}, {"socrates": 0.02849802310899971, "ibn_sina": 0.008179125116102977, "hypatia": 0.00619593587554379, "einstein": 0.022452466805267794, "sagan": 0.916251663504158, "darwin": 0.018422785589927618}], "prompt_tokens": 523, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.04232267771053761, "einstein": 0.025590383371811, "hypatia": 0.007144521263875355, "ibn_sina": 0.010485440165061352, "sagan": 0.8908750666343173, "socrates": 0.023581910854397307}, "confidence": 0.7561858675353944}}, "usage": {"input_tokens": 523, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.010030933232457169, "hypatia": 0.008423258504315159, "sagan": 0.9230459027590482, "socrates": 0.02646482751432718, "darwin": 0.015553360737548604, "einstein": 0.016481717252303695}, "confidence": 0.8240791007617492}}, "usage": {"input_tokens": 523, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.007279833257480972, "darwin": 0.015912985093585696, "sagan": 0.9046288193968632, "ibn_sina": 0.010990038830519813, "einstein": 0.018881356198795482, "socrates": 0.04230696722275484}, "confidence": 0.7851120231894368}}, "usage": {"input_tokens": 523, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.011307818645799855, "einstein": 0.023646815020430846, "darwin": 0.017448949889511577, "sagan": 0.9116520150787981, "hypatia": 0.009738532068183972, "socrates": 0.026205869297275718}, "confidence": 0.7994589851981025}}, "usage": {"input_tokens": 523, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.01715740774084351, "sagan": 0.9301635774951261, "socrates": 0.027046784479573635, "ibn_sina": 0.00889261518480956, "darwin": 0.010499083584899593, "hypatia": 0.0062405315147475065}, "confidence": 0.8397501276209631}}, "usage": {"input_tokens": 523, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.02849802310899971, "ibn_sina": 0.008179125116102977, "hypatia": 0.00619593587554379, "einstein": 0.022452466805267794, "sagan": 0.916251663504158, "darwin": 0.018422785589927618}, "confidence": 0.8095336578385967}}, "usage": {"input_tokens": 523, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:51Z"}
{"id": "c-sagan-01-dessert-01", "uid": "c074", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "prompt_sha256_by_order": ["5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d"], "choice": "sagan", "probs": {"darwin": 0.01216061422534042, "einstein": 0.02587826842379306, "hypatia": 0.02523310518944203, "ibn_sina": 0.01889480644898629, "sagan": 0.8960148390316276, "socrates": 0.02181836668081061}, "confidence": 0.7661559058152273, "seconds": 0.1920689841666293, "seconds_all_orders": 1.1524139049997757, "seconds_by_order": [0.19179555200025789, 0.19203253900013806, 0.19225550800001656, 0.19211781900048663, 0.1921852179984853, 0.19202726900039124], "argmax_stable": true, "spread": {"darwin": 0.006234819690907514, "einstein": 0.005307419996244719, "hypatia": 0.01029251644319712, "ibn_sina": 0.01079137800481724, "sagan": 0.028722191765262917, "socrates": 0.021864979749553294}, "spread_2dp": {"darwin": 0.01, "einstein": 0.0, "hypatia": 0.009999999999999998, "ibn_sina": 0.01, "sagan": 0.020000000000000018, "socrates": 0.019999999999999997}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.01216061422534042, "einstein": 0.02587826842379306, "hypatia": 0.02523310518944203, "ibn_sina": 0.01889480644898629, "sagan": 0.8960148390316276, "socrates": 0.02181836668081061}, {"ibn_sina": 0.01194941795991138, "hypatia": 0.027358939557804288, "sagan": 0.9010172584372659, "socrates": 0.01890624188975891, "darwin": 0.012099109105489874, "einstein": 0.0286690330497697}, {"hypatia": 0.017066423114607168, "darwin": 0.012464925356624947, "sagan": 0.8987399182727366, "ibn_sina": 0.013599357901956056, "einstein": 0.02604040899536368, "socrates": 0.03208896635871151}, {"ibn_sina": 0.014365847674301782, "einstein": 0.029338552204811328, "darwin": 0.015124665669679578, "sagan": 0.899659811694288, "hypatia": 0.023903884467014885, "socrates": 0.017607238289904403}, {"einstein": 0.030618628433908994, "sagan": 0.9150567149109097, "socrates": 0.016201153865851945, "ibn_sina": 0.010954098384344703, "darwin": 0.008889845978772064, "hypatia": 0.01827955842621266}, {"socrates": 0.010223986609158216, "ibn_sina": 0.00810342844416905, "hypatia": 0.0183410629282449, "einstein": 0.025311208437664275, "sagan": 0.9247370307968905, "darwin": 0.01328328278387311}], "prompt_tokens": 483, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.01216061422534042, "einstein": 0.02587826842379306, "hypatia": 0.02523310518944203, "ibn_sina": 0.01889480644898629, "sagan": 0.8960148390316276, "socrates": 0.02181836668081061}, "confidence": 0.7661559058152273}}, "usage": {"input_tokens": 483, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.01194941795991138, "hypatia": 0.027358939557804288, "sagan": 0.9010172584372659, "socrates": 0.01890624188975891, "darwin": 0.012099109105489874, "einstein": 0.0286690330497697}, "confidence": 0.7768589776535972}}, "usage": {"input_tokens": 483, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.017066423114607168, "darwin": 0.012464925356624947, "sagan": 0.8987399182727366, "ibn_sina": 0.013599357901956056, "einstein": 0.02604040899536368, "socrates": 0.03208896635871151}, "confidence": 0.7720873900683042}}, "usage": {"input_tokens": 483, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.014365847674301782, "einstein": 0.029338552204811328, "darwin": 0.015124665669679578, "sagan": 0.899659811694288, "hypatia": 0.023903884467014885, "socrates": 0.017607238289904403}, "confidence": 0.7738780852565282}}, "usage": {"input_tokens": 483, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.030618628433908994, "sagan": 0.9150567149109097, "socrates": 0.016201153865851945, "ibn_sina": 0.010954098384344703, "darwin": 0.008889845978772064, "hypatia": 0.01827955842621266}, "confidence": 0.8068743198238024}}, "usage": {"input_tokens": 483, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.010223986609158216, "ibn_sina": 0.00810342844416905, "hypatia": 0.0183410629282449, "einstein": 0.025311208437664275, "sagan": 0.9247370307968905, "darwin": 0.01328328278387311}, "confidence": 0.8277547228542547}}, "usage": {"input_tokens": 483, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:52Z"}
{"id": "c-sagan-01-dessert-03", "uid": "c075", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "prompt_sha256_by_order": ["d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d"], "choice": "sagan", "probs": {"darwin": 0.003425082930715437, "einstein": 0.029655101770336548, "hypatia": 0.003420802581969342, "ibn_sina": 0.0033121937195606946, "sagan": 0.9554388145296371, "socrates": 0.004748004467781015}, "confidence": 0.8965596407534276, "seconds": 0.21942223716693357, "seconds_all_orders": 1.3165334230016015, "seconds_by_order": [0.21924780500012275, 0.2195378830001573, 0.21930880500076455, 0.2196848420007882, 0.21928650500012736, 0.2194675829996413], "argmax_stable": true, "spread": {"darwin": 0.0025584554935028396, "einstein": 0.03420168723500022, "hypatia": 0.0019874319203212996, "ibn_sina": 0.0021805038090521346, "sagan": 0.03874191657644832, "socrates": 0.0039611467669459725}, "spread_2dp": {"darwin": 0.01, "einstein": 0.039999999999999994, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.039999999999999925, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.003425082930715437, "einstein": 0.029655101770336548, "hypatia": 0.003420802581969342, "ibn_sina": 0.0033121937195606946, "sagan": 0.9554388145296371, "socrates": 0.004748004467781015}, {"ibn_sina": 0.0035808873919134666, "hypatia": 0.005408234502290642, "sagan": 0.9239915138818267, "socrates": 0.006191459962353791, "darwin": 0.004505202875584179, "einstein": 0.056322701386031314}, {"hypatia": 0.004331217246665624, "darwin": 0.004420197043779741, "sagan": 0.9447401908417351, "ibn_sina": 0.004121103622135952, "einstein": 0.03737104870461271, "socrates": 0.00501624254107089}, {"ibn_sina": 0.0043656344765439295, "einstein": 0.03707617143176732, "darwin": 0.005016290175923022, "sagan": 0.9445933384098553, "hypatia": 0.004809956343862672, "socrates": 0.004138609162047746}, {"einstein": 0.022121014151031095, "sagan": 0.962733430458275, "socrates": 0.00644289417667875, "ibn_sina": 0.0023917943416326345, "darwin": 0.0024578346824201827, "hypatia": 0.003853032189962378}, {"socrates": 0.0024817474097327775, "ibn_sina": 0.002185130667491795, "hypatia": 0.0036820262284135786, "einstein": 0.03149983758009509, "sagan": 0.9559555956315353, "darwin": 0.004195662482731447}], "prompt_tokens": 561, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.003425082930715437, "einstein": 0.029655101770336548, "hypatia": 0.003420802581969342, "ibn_sina": 0.0033121937195606946, "sagan": 0.9554388145296371, "socrates": 0.004748004467781015}, "confidence": 0.8965596407534276}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.0035808873919134666, "hypatia": 0.005408234502290642, "sagan": 0.9239915138818267, "socrates": 0.006191459962353791, "darwin": 0.004505202875584179, "einstein": 0.056322701386031314}, "confidence": 0.828439920641693}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.004331217246665624, "darwin": 0.004420197043779741, "sagan": 0.9447401908417351, "ibn_sina": 0.004121103622135952, "einstein": 0.03737104870461271, "socrates": 0.00501624254107089}, "confidence": 0.8728132806903993}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.0043656344765439295, "einstein": 0.03707617143176732, "darwin": 0.005016290175923022, "sagan": 0.9445933384098553, "hypatia": 0.004809956343862672, "socrates": 0.004138609162047746}, "confidence": 0.8724588437843813}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.022121014151031095, "sagan": 0.962733430458275, "socrates": 0.00644289417667875, "ibn_sina": 0.0023917943416326345, "darwin": 0.0024578346824201827, "hypatia": 0.003853032189962378}, "confidence": 0.912895738915547}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.0024817474097327775, "ibn_sina": 0.002185130667491795, "hypatia": 0.0036820262284135786, "einstein": 0.03149983758009509, "sagan": 0.9559555956315353, "darwin": 0.004195662482731447}, "confidence": 0.8978625224246457}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:54Z"}
{"id": "c-sagan-01-dessert-03", "uid": "c076", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "prompt_sha256_by_order": ["16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac"], "choice": "sagan", "probs": {"darwin": 0.011742834994669014, "einstein": 0.09060783484392368, "hypatia": 0.008979695877631998, "ibn_sina": 0.02294494348310091, "sagan": 0.847284265361911, "socrates": 0.018440425438763487}, "confidence": 0.672622545878259, "seconds": 0.25678591966667835, "seconds_all_orders": 1.54071551800007, "seconds_by_order": [0.25637764900056936, 0.2573155019999831, 0.2564856890003284, 0.25707102399974247, 0.25617263200001616, 0.2572930219994305], "argmax_stable": true, "spread": {"darwin": 0.004928095617925102, "einstein": 0.07677917084755279, "hypatia": 0.003020026718302509, "ibn_sina": 0.009236231328045953, "sagan": 0.08226658120020669, "socrates": 0.010531968433651461}, "spread_2dp": {"darwin": 0.0, "einstein": 0.08000000000000002, "hypatia": 0.0, "ibn_sina": 0.009999999999999998, "sagan": 0.08999999999999997, "socrates": 0.009999999999999998}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.011742834994669014, "einstein": 0.09060783484392368, "hypatia": 0.008979695877631998, "ibn_sina": 0.02294494348310091, "sagan": 0.847284265361911, "socrates": 0.018440425438763487}, {"ibn_sina": 0.02505801830844419, "hypatia": 0.01057122029814999, "sagan": 0.8498244262237231, "socrates": 0.017582236972353414, "darwin": 0.012457482815139413, "einstein": 0.0845066153821899}, {"hypatia": 0.009798696064442673, "darwin": 0.011456779240261759, "sagan": 0.8027791178182438, "ibn_sina": 0.01868074381188633, "einstein": 0.1394297501710613, "socrates": 0.017854912894104104}, {"ibn_sina": 0.021151656648355967, "einstein": 0.10060969538391937, "darwin": 0.013919848126497146, "sagan": 0.8300459709179748, "hypatia": 0.008405455558164064, "socrates": 0.02586737336508868}, {"einstein": 0.06265057932350851, "sagan": 0.8850456990184505, "socrates": 0.019328098609852996, "ibn_sina": 0.016432676959768383, "darwin": 0.008991752508572044, "hypatia": 0.007551193579847481}, {"socrates": 0.01533540493143722, "ibn_sina": 0.01582178698039824, "hypatia": 0.009658112909965249, "einstein": 0.11934325430583012, "sagan": 0.8266878950610346, "darwin": 0.013153545811334542}], "prompt_tokens": 721, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.011742834994669014, "einstein": 0.09060783484392368, "hypatia": 0.008979695877631998, "ibn_sina": 0.02294494348310091, "sagan": 0.847284265361911, "socrates": 0.018440425438763487}, "confidence": 0.672622545878259}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.02505801830844419, "hypatia": 0.01057122029814999, "sagan": 0.8498244262237231, "socrates": 0.017582236972353414, "darwin": 0.012457482815139413, "einstein": 0.0845066153821899}, "confidence": 0.6766562828380162}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.009798696064442673, "darwin": 0.011456779240261759, "sagan": 0.8027791178182438, "ibn_sina": 0.01868074381188633, "einstein": 0.1394297501710613, "socrates": 0.017854912894104104}, "confidence": 0.5977480090920967}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.021151656648355967, "einstein": 0.10060969538391937, "darwin": 0.013919848126497146, "sagan": 0.8300459709179748, "hypatia": 0.008405455558164064, "socrates": 0.02586737336508868}, "confidence": 0.6405754624974309}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.06265057932350851, "sagan": 0.8850456990184505, "socrates": 0.019328098609852996, "ibn_sina": 0.016432676959768383, "darwin": 0.008991752508572044, "hypatia": 0.007551193579847481}, "confidence": 0.7456149578154652}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.01533540493143722, "ibn_sina": 0.01582178698039824, "hypatia": 0.009658112909965249, "einstein": 0.11934325430583012, "sagan": 0.8266878950610346, "darwin": 0.013153545811334542}, "confidence": 0.6380889840265682}}, "usage": {"input_tokens": 721, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:55Z"}
{"id": "c-sagan-01-dessert-03", "uid": "c077", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "prompt_sha256_by_order": ["6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a"], "choice": "sagan", "probs": {"darwin": 0.016693618651787118, "einstein": 0.010030544299005234, "hypatia": 0.054667248644651154, "ibn_sina": 0.06486943916553849, "sagan": 0.7943972732617883, "socrates": 0.05934187597722965}, "confidence": 0.5705971923338975, "seconds": 0.21970521983378907, "seconds_all_orders": 1.3182313190027344, "seconds_by_order": [0.21966399200027809, 0.2198036610006966, 0.2193466240005364, 0.21955994200106943, 0.2203050570005871, 0.21955204299956677], "argmax_stable": true, "spread": {"darwin": 0.007245680473296677, "einstein": 0.00821069802375154, "hypatia": 0.09449695810083864, "ibn_sina": 0.04069250489820805, "sagan": 0.08707444354621408, "socrates": 0.10716818256484874}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.1, "ibn_sina": 0.039999999999999994, "sagan": 0.08000000000000007, "socrates": 0.11}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.016693618651787118, "einstein": 0.010030544299005234, "hypatia": 0.054667248644651154, "ibn_sina": 0.06486943916553849, "sagan": 0.7943972732617883, "socrates": 0.05934187597722965}, {"ibn_sina": 0.04593593669781261, "hypatia": 0.11668331975945315, "sagan": 0.790312760125401, "socrates": 0.020545120560690897, "darwin": 0.010542154577004081, "einstein": 0.015980708279638305}, {"hypatia": 0.08797886048933487, "darwin": 0.010977894134319297, "sagan": 0.8059665675985777, "ibn_sina": 0.02417693426733044, "einstein": 0.011976849017382322, "socrates": 0.058922894493055496}, {"ibn_sina": 0.04070794756298732, "einstein": 0.012254496442467344, "darwin": 0.017607945869765402, "sagan": 0.7253646934425491, "hypatia": 0.07635161355669119, "socrates": 0.12771330312553963}, {"einstein": 0.017463373129917317, "sagan": 0.765558891373659, "socrates": 0.03434329659466474, "ibn_sina": 0.03563607647987186, "darwin": 0.010610225238971177, "hypatia": 0.13638813718291595}, {"socrates": 0.07703571971838016, "ibn_sina": 0.032604886837721815, "hypatia": 0.041891179082077325, "einstein": 0.018241242322756774, "sagan": 0.8124391369887631, "darwin": 0.01778783505030076}], "prompt_tokens": 565, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.016693618651787118, "einstein": 0.010030544299005234, "hypatia": 0.054667248644651154, "ibn_sina": 0.06486943916553849, "sagan": 0.7943972732617883, "socrates": 0.05934187597722965}, "confidence": 0.5705971923338975}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.04593593669781261, "hypatia": 0.11668331975945315, "sagan": 0.790312760125401, "socrates": 0.020545120560690897, "darwin": 0.010542154577004081, "einstein": 0.015980708279638305}, "confidence": 0.5693295858958234}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.08797886048933487, "darwin": 0.010977894134319297, "sagan": 0.8059665675985777, "ibn_sina": 0.02417693426733044, "einstein": 0.011976849017382322, "socrates": 0.058922894493055496}, "confidence": 0.5939713344371806}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.04070794756298732, "einstein": 0.012254496442467344, "darwin": 0.017607945869765402, "sagan": 0.7253646934425491, "hypatia": 0.07635161355669119, "socrates": 0.12771330312553963}, "confidence": 0.46049385353986166}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.017463373129917317, "sagan": 0.765558891373659, "socrates": 0.03434329659466474, "ibn_sina": 0.03563607647987186, "darwin": 0.010610225238971177, "hypatia": 0.13638813718291595}, "confidence": 0.5290588940488677}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.07703571971838016, "ibn_sina": 0.032604886837721815, "hypatia": 0.041891179082077325, "einstein": 0.018241242322756774, "sagan": 0.8124391369887631, "darwin": 0.01778783505030076}, "confidence": 0.6033507435417}}, "usage": {"input_tokens": 565, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:57Z"}
{"id": "c-sagan-01-dessert-03", "uid": "c078", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "prompt_sha256_by_order": ["2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541"], "choice": "sagan", "probs": {"darwin": 0.011826826898256874, "einstein": 0.017548017783976147, "hypatia": 0.004137965544553237, "ibn_sina": 0.005486413697042867, "sagan": 0.9544792318615173, "socrates": 0.006521544214653678}, "confidence": 0.8938817978207523, "seconds": 0.22533269899940933, "seconds_all_orders": 1.351996193996456, "seconds_by_order": [0.2252200869988883, 0.22526315600043745, 0.22593040100036887, 0.22492998899906524, 0.2257668719994399, 0.22488568899825623], "argmax_stable": true, "spread": {"darwin": 0.006694198130327412, "einstein": 0.011303540015803635, "hypatia": 0.002524196420747119, "ibn_sina": 0.004271280550700079, "sagan": 0.017209988729465775, "socrates": 0.00347011422505703}, "spread_2dp": {"darwin": 0.0, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.010000000000000009, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.011826826898256874, "einstein": 0.017548017783976147, "hypatia": 0.004137965544553237, "ibn_sina": 0.005486413697042867, "sagan": 0.9544792318615173, "socrates": 0.006521544214653678}, {"ibn_sina": 0.006343696525880554, "hypatia": 0.006278238595775887, "sagan": 0.9470422723483083, "socrates": 0.006644097024590161, "darwin": 0.009308046750941264, "einstein": 0.024383648754503946}, {"hypatia": 0.0060955258469494525, "darwin": 0.008985643707767856, "sagan": 0.9512187565555346, "ibn_sina": 0.006380531078036711, "einstein": 0.019841217525173305, "socrates": 0.0074783252865380755}, {"ibn_sina": 0.007880832841113353, "einstein": 0.013080108738700311, "darwin": 0.008546276918579954, "sagan": 0.9578632535966023, "hypatia": 0.005898887519489928, "socrates": 0.006730640385514089}, {"einstein": 0.016227644105744926, "sagan": 0.9640830552407399, "socrates": 0.005315028553459361, "ibn_sina": 0.004168687571257306, "darwin": 0.0051326287679294615, "hypatia": 0.005072955760868941}, {"socrates": 0.004008211061481045, "ibn_sina": 0.0036095522904132744, "hypatia": 0.0037540421750287684, "einstein": 0.01493161031576137, "sagan": 0.9642522610777741, "darwin": 0.009444323079541484}], "prompt_tokens": 625, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.011826826898256874, "einstein": 0.017548017783976147, "hypatia": 0.004137965544553237, "ibn_sina": 0.005486413697042867, "sagan": 0.9544792318615173, "socrates": 0.006521544214653678}, "confidence": 0.8938817978207523}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.006343696525880554, "hypatia": 0.006278238595775887, "sagan": 0.9470422723483083, "socrates": 0.006644097024590161, "darwin": 0.009308046750941264, "einstein": 0.024383648754503946}, "confidence": 0.8772328845595577}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.0060955258469494525, "darwin": 0.008985643707767856, "sagan": 0.9512187565555346, "ibn_sina": 0.006380531078036711, "einstein": 0.019841217525173305, "socrates": 0.0074783252865380755}, "confidence": 0.8865103965880738}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.007880832841113353, "einstein": 0.013080108738700311, "darwin": 0.008546276918579954, "sagan": 0.9578632535966023, "hypatia": 0.005898887519489928, "socrates": 0.006730640385514089}, "confidence": 0.9014660159256286}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.016227644105744926, "sagan": 0.9640830552407399, "socrates": 0.005315028553459361, "ibn_sina": 0.004168687571257306, "darwin": 0.0051326287679294615, "hypatia": 0.005072955760868941}, "confidence": 0.9157806160940251}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.004008211061481045, "ibn_sina": 0.0036095522904132744, "hypatia": 0.0037540421750287684, "einstein": 0.01493161031576137, "sagan": 0.9642522610777741, "darwin": 0.009444323079541484}, "confidence": 0.9161653104099458}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:58Z"}
{"id": "c-sagan-01-dessert-03", "uid": "c079", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "prompt_sha256_by_order": ["fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520"], "choice": "sagan", "probs": {"darwin": 0.07946197165471632, "einstein": 0.029907371943454455, "hypatia": 0.015331737464559857, "ibn_sina": 0.01330698860108526, "sagan": 0.7968169814590469, "socrates": 0.06517494887713719}, "confidence": 0.5761430438293302, "seconds": 0.25456512783330254, "seconds_all_orders": 1.5273907669998152, "seconds_by_order": [0.25476785299906624, 0.25472966300003463, 0.2540299789998244, 0.25402137899982336, 0.2549591610004427, 0.25488273200062395], "argmax_stable": true, "spread": {"darwin": 0.032908335962914564, "einstein": 0.011243926018539685, "hypatia": 0.03529344286388077, "ibn_sina": 0.005296217995904861, "sagan": 0.1275438375355794, "socrates": 0.11684447268921704}, "spread_2dp": {"darwin": 0.03, "einstein": 0.009999999999999998, "hypatia": 0.030000000000000002, "ibn_sina": 0.01, "sagan": 0.13, "socrates": 0.12}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.07946197165471632, "einstein": 0.029907371943454455, "hypatia": 0.015331737464559857, "ibn_sina": 0.01330698860108526, "sagan": 0.7968169814590469, "socrates": 0.06517494887713719}, {"ibn_sina": 0.01860320659699012, "hypatia": 0.029875380360136074, "sagan": 0.7964469336896118, "socrates": 0.0382415377578173, "darwin": 0.08244831942811043, "einstein": 0.03438462216733415}, {"hypatia": 0.05062518032844062, "darwin": 0.057007247181862106, "sagan": 0.7246068334135021, "ibn_sina": 0.014298135069813407, "einstein": 0.030881914671906208, "socrates": 0.12258068933447566}, {"ibn_sina": 0.013706088335729387, "einstein": 0.025083455441665112, "darwin": 0.0598927356537222, "sagan": 0.7780058892385957, "hypatia": 0.02074058712259284, "socrates": 0.10257124420769478}, {"einstein": 0.023140696148794465, "sagan": 0.8377078937699706, "socrates": 0.03426191299185029, "ibn_sina": 0.01715773953474607, "darwin": 0.049539983465195864, "hypatia": 0.03819177408944267}, {"socrates": 0.15110638568106732, "ibn_sina": 0.014436116081130036, "hypatia": 0.020570321800360294, "einstein": 0.032230168333435845, "sagan": 0.7101640562343912, "darwin": 0.07149295186961531}], "prompt_tokens": 697, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.07946197165471632, "einstein": 0.029907371943454455, "hypatia": 0.015331737464559857, "ibn_sina": 0.01330698860108526, "sagan": 0.7968169814590469, "socrates": 0.06517494887713719}, "confidence": 0.5761430438293302}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.01860320659699012, "hypatia": 0.029875380360136074, "sagan": 0.7964469336896118, "socrates": 0.0382415377578173, "darwin": 0.08244831942811043, "einstein": 0.03438462216733415}, "confidence": 0.5740105343909567}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.05062518032844062, "darwin": 0.057007247181862106, "sagan": 0.7246068334135021, "ibn_sina": 0.014298135069813407, "einstein": 0.030881914671906208, "socrates": 0.12258068933447566}, "confidence": 0.4564623434344918}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.013706088335729387, "einstein": 0.025083455441665112, "darwin": 0.0598927356537222, "sagan": 0.7780058892385957, "hypatia": 0.02074058712259284, "socrates": 0.10257124420769478}, "confidence": 0.5447780465930678}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.023140696148794465, "sagan": 0.8377078937699706, "socrates": 0.03426191299185029, "ibn_sina": 0.01715773953474607, "darwin": 0.049539983465195864, "hypatia": 0.03819177408944267}, "confidence": 0.6492053144564863}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.15110638568106732, "ibn_sina": 0.014436116081130036, "hypatia": 0.020570321800360294, "einstein": 0.032230168333435845, "sagan": 0.7101640562343912, "darwin": 0.07149295186961531}, "confidence": 0.4407372304782278}}, "usage": {"input_tokens": 697, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:18:59Z"}
{"id": "c-sagan-01-main-01", "uid": "c080", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "prompt_sha256_by_order": ["140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631"], "choice": "sagan", "probs": {"darwin": 0.018424649638427177, "einstein": 0.05277379041437551, "hypatia": 0.022272226383284546, "ibn_sina": 0.014743701211146917, "sagan": 0.8613980292581698, "socrates": 0.030387603094596018}, "confidence": 0.6961215288329974, "seconds": 0.22041484933348934, "seconds_all_orders": 1.3224890960009361, "seconds_by_order": [0.22022836799988, 0.22015268799987098, 0.22064847399997234, 0.22070156300105737, 0.2207320739998977, 0.22002592900025775], "argmax_stable": true, "spread": {"darwin": 0.022238163591535975, "einstein": 0.02785440666882598, "hypatia": 0.004874732533590453, "ibn_sina": 0.006728197508552306, "sagan": 0.03980964317506874, "socrates": 0.02363358295359933}, "spread_2dp": {"darwin": 0.02, "einstein": 0.030000000000000006, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.040000000000000036, "socrates": 0.020000000000000004}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.018424649638427177, "einstein": 0.05277379041437551, "hypatia": 0.022272226383284546, "ibn_sina": 0.014743701211146917, "sagan": 0.8613980292581698, "socrates": 0.030387603094596018}, {"ibn_sina": 0.011128868558469765, "hypatia": 0.023146055939288458, "sagan": 0.8436674671680058, "socrates": 0.030371523375602352, "darwin": 0.04066281322996315, "einstein": 0.051023271728670425}, {"hypatia": 0.024011370271661718, "darwin": 0.025956818785941076, "sagan": 0.823494182040483, "ibn_sina": 0.01710276751337557, "einstein": 0.05811053481088525, "socrates": 0.051324326577653435}, {"ibn_sina": 0.014416853795196227, "einstein": 0.04867528574973222, "darwin": 0.03031260434857233, "sagan": 0.8565796018835508, "hypatia": 0.020644293317148538, "socrates": 0.02937136090579981}, {"einstein": 0.04306181813463901, "sagan": 0.8633038252155517, "socrates": 0.027690743624054104, "ibn_sina": 0.014579832137658648, "darwin": 0.02689985440267554, "hypatia": 0.024463926485420905}, {"socrates": 0.03337926683724579, "ibn_sina": 0.010374570004823263, "hypatia": 0.019589193951830453, "einstein": 0.07091622480346499, "sagan": 0.8378049917401231, "darwin": 0.02793575266251234}], "prompt_tokens": 573, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.018424649638427177, "einstein": 0.05277379041437551, "hypatia": 0.022272226383284546, "ibn_sina": 0.014743701211146917, "sagan": 0.8613980292581698, "socrates": 0.030387603094596018}, "confidence": 0.6961215288329974}}, "usage": {"input_tokens": 573, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.011128868558469765, "hypatia": 0.023146055939288458, "sagan": 0.8436674671680058, "socrates": 0.030371523375602352, "darwin": 0.04066281322996315, "einstein": 0.051023271728670425}, "confidence": 0.6611363858181303}}, "usage": {"input_tokens": 573, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.024011370271661718, "darwin": 0.025956818785941076, "sagan": 0.823494182040483, "ibn_sina": 0.01710276751337557, "einstein": 0.05811053481088525, "socrates": 0.051324326577653435}, "confidence": 0.6228357947315277}}, "usage": {"input_tokens": 573, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.014416853795196227, "einstein": 0.04867528574973222, "darwin": 0.03031260434857233, "sagan": 0.8565796018835508, "hypatia": 0.020644293317148538, "socrates": 0.02937136090579981}, "confidence": 0.6862161533794826}}, "usage": {"input_tokens": 573, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.04306181813463901, "sagan": 0.8633038252155517, "socrates": 0.027690743624054104, "ibn_sina": 0.014579832137658648, "darwin": 0.02689985440267554, "hypatia": 0.024463926485420905}, "confidence": 0.699339099359739}}, "usage": {"input_tokens": 573, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.03337926683724579, "ibn_sina": 0.010374570004823263, "hypatia": 0.019589193951830453, "einstein": 0.07091622480346499, "sagan": 0.8378049917401231, "darwin": 0.02793575266251234}, "confidence": 0.6511987180947086}}, "usage": {"input_tokens": 573, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:01Z"}
{"id": "c-sagan-01-main-02", "uid": "c081", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "prompt_sha256_by_order": ["365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3"], "choice": "sagan", "probs": {"darwin": 0.011082608170756318, "einstein": 0.15429258611300517, "hypatia": 0.007066749842143682, "ibn_sina": 0.007917427557988056, "sagan": 0.8033026741166169, "socrates": 0.016337954199489885}, "confidence": 0.6035245191233265, "seconds": 0.2266048450001108, "seconds_all_orders": 1.3596290700006648, "seconds_by_order": [0.22690153300027305, 0.22665042500011623, 0.22745873799976835, 0.22574192200045218, 0.22645390599973325, 0.22642254600032174], "argmax_stable": true, "spread": {"darwin": 0.006232110055405915, "einstein": 0.09268683665543026, "hypatia": 0.002985170923648253, "ibn_sina": 0.003267393748497649, "sagan": 0.08223240433017409, "socrates": 0.009788960801900502}, "spread_2dp": {"darwin": 0.0, "einstein": 0.09000000000000001, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.08999999999999997, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.011082608170756318, "einstein": 0.15429258611300517, "hypatia": 0.007066749842143682, "ibn_sina": 0.007917427557988056, "sagan": 0.8033026741166169, "socrates": 0.016337954199489885}, {"ibn_sina": 0.006217282149399697, "hypatia": 0.008137496552872877, "sagan": 0.8431800904086303, "socrates": 0.015442924565921124, "darwin": 0.014796353543129796, "einstein": 0.11222585278004628}, {"hypatia": 0.005655706473236945, "darwin": 0.01037763469822579, "sagan": 0.7770342219701544, "ibn_sina": 0.008712062665128698, "einstein": 0.17617373311566678, "socrates": 0.02204664107758738}, {"ibn_sina": 0.008277332030193076, "einstein": 0.11053294509496744, "darwin": 0.012711903876574025, "sagan": 0.846119724118451, "hypatia": 0.007681353218709344, "socrates": 0.014676741661105105}, {"einstein": 0.2032197817503977, "sagan": 0.763887319788277, "socrates": 0.01373166042774583, "ibn_sina": 0.0054446689166310494, "darwin": 0.008564243487723881, "hypatia": 0.005152325629224624}, {"socrates": 0.02352062122964633, "ibn_sina": 0.007128956676690482, "hypatia": 0.006339917827933781, "einstein": 0.12150290836002729, "sagan": 0.8303668836934608, "darwin": 0.011140712212241327}], "prompt_tokens": 637, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.011082608170756318, "einstein": 0.15429258611300517, "hypatia": 0.007066749842143682, "ibn_sina": 0.007917427557988056, "sagan": 0.8033026741166169, "socrates": 0.016337954199489885}, "confidence": 0.6035245191233265}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.006217282149399697, "hypatia": 0.008137496552872877, "sagan": 0.8431800904086303, "socrates": 0.015442924565921124, "darwin": 0.014796353543129796, "einstein": 0.11222585278004628}, "confidence": 0.6689315156062119}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.005655706473236945, "darwin": 0.01037763469822579, "sagan": 0.7770342219701544, "ibn_sina": 0.008712062665128698, "einstein": 0.17617373311566678, "socrates": 0.02204664107758738}, "confidence": 0.5626252037067996}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.008277332030193076, "einstein": 0.11053294509496744, "darwin": 0.012711903876574025, "sagan": 0.846119724118451, "hypatia": 0.007681353218709344, "socrates": 0.014676741661105105}, "confidence": 0.6743687633828962}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.2032197817503977, "sagan": 0.763887319788277, "socrates": 0.01373166042774583, "ibn_sina": 0.0054446689166310494, "darwin": 0.008564243487723881, "hypatia": 0.005152325629224624}, "confidence": 0.5501682552059414}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.02352062122964633, "ibn_sina": 0.007128956676690482, "hypatia": 0.006339917827933781, "einstein": 0.12150290836002729, "sagan": 0.8303668836934608, "darwin": 0.011140712212241327}, "confidence": 0.6460485639372667}}, "usage": {"input_tokens": 637, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:02Z"}
{"id": "c-sagan-01-main-02", "uid": "c082", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "prompt_sha256_by_order": ["dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc"], "choice": "einstein", "probs": {"darwin": 0.03161843471527042, "einstein": 0.8202299262310274, "hypatia": 0.029954503277019935, "ibn_sina": 0.024226056887533362, "sagan": 0.05328807546977872, "socrates": 0.04068300341937018}, "confidence": 0.6157069085826219, "seconds": 0.19003406716637983, "seconds_all_orders": 1.140204402998279, "seconds_by_order": [0.1906025509997562, 0.18969364799886534, 0.18999158599945076, 0.18992185599927325, 0.18951175999973202, 0.19048300200120138], "argmax_stable": true, "spread": {"darwin": 0.01646419735008799, "einstein": 0.04729155534865226, "hypatia": 0.014243298918869558, "ibn_sina": 0.0159747157068248, "sagan": 0.022956389492380685, "socrates": 0.022493768972316246}, "spread_2dp": {"darwin": 0.010000000000000002, "einstein": 0.04999999999999993, "hypatia": 0.010000000000000002, "ibn_sina": 0.02, "sagan": 0.020000000000000004, "socrates": 0.019999999999999997}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.03161843471527042, "einstein": 0.8202299262310274, "hypatia": 0.029954503277019935, "ibn_sina": 0.024226056887533362, "sagan": 0.05328807546977872, "socrates": 0.04068300341937018}, {"ibn_sina": 0.031206518886856202, "hypatia": 0.0324848392281738, "sagan": 0.04844697275054109, "socrates": 0.04150082812095457, "darwin": 0.042363607218004404, "einstein": 0.80399723379547}, {"hypatia": 0.041000826606876054, "darwin": 0.03477898515701122, "sagan": 0.04792563830178284, "ibn_sina": 0.04017940666026823, "einstein": 0.7729383708823752, "socrates": 0.06317677239168643}, {"ibn_sina": 0.0399532286703146, "einstein": 0.785154136032969, "darwin": 0.03477518241805623, "sagan": 0.05035265264226081, "hypatia": 0.04419780219588949, "socrates": 0.04556699804050984}, {"einstein": 0.7951963174685563, "sagan": 0.07088202779416353, "socrates": 0.04304245621550358, "ibn_sina": 0.02420469095344343, "darwin": 0.03411666415813317, "hypatia": 0.03255784341019987}, {"socrates": 0.057003548793117934, "ibn_sina": 0.032478446603713636, "hypatia": 0.03688893318781068, "einstein": 0.7796753351262566, "sagan": 0.06805432642118467, "darwin": 0.025899409867916414}], "prompt_tokens": 537, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.03161843471527042, "einstein": 0.8202299262310274, "hypatia": 0.029954503277019935, "ibn_sina": 0.024226056887533362, "sagan": 0.05328807546977872, "socrates": 0.04068300341937018}, "confidence": 0.6157069085826219}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.031206518886856202, "hypatia": 0.0324848392281738, "sagan": 0.04844697275054109, "socrates": 0.04150082812095457, "darwin": 0.042363607218004404, "einstein": 0.80399723379547}, "confidence": 0.585165720005706}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.041000826606876054, "darwin": 0.03477898515701122, "sagan": 0.04792563830178284, "ibn_sina": 0.04017940666026823, "einstein": 0.7729383708823752, "socrates": 0.06317677239168643}, "confidence": 0.5298723122432012}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.0399532286703146, "einstein": 0.785154136032969, "darwin": 0.03477518241805623, "sagan": 0.05035265264226081, "hypatia": 0.04419780219588949, "socrates": 0.04556699804050984}, "confidence": 0.5510053333366629}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.7951963174685563, "sagan": 0.07088202779416353, "socrates": 0.04304245621550358, "ibn_sina": 0.02420469095344343, "darwin": 0.03411666415813317, "hypatia": 0.03255784341019987}, "confidence": 0.5704287102661393}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.057003548793117934, "ibn_sina": 0.032478446603713636, "hypatia": 0.03688893318781068, "einstein": 0.7796753351262566, "sagan": 0.06805432642118467, "darwin": 0.025899409867916414}, "confidence": 0.542633015729107}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:03Z"}
{"id": "c-sagan-01-main-03", "uid": "c083", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "prompt_sha256_by_order": ["8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad"], "choice": "sagan", "probs": {"darwin": 0.015242668900582453, "einstein": 0.17220471693143607, "hypatia": 0.011709092132953687, "ibn_sina": 0.02172903688330456, "sagan": 0.765968549482974, "socrates": 0.013145935668749116}, "confidence": 0.540852030151313, "seconds": 0.2546663254994807, "seconds_all_orders": 1.5279979529968841, "seconds_by_order": [0.25549033700008295, 0.254249556999639, 0.25420661799944355, 0.25526441900001373, 0.2548279429993272, 0.25395907899837766], "argmax_stable": true, "spread": {"darwin": 0.005561695562778, "einstein": 0.13759190561123585, "hypatia": 0.00474552251591397, "ibn_sina": 0.005042244566025752, "sagan": 0.12444674643709841, "socrates": 0.0037913152435209464}, "spread_2dp": {"darwin": 0.01, "einstein": 0.14, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.13, "socrates": 0.0}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.015242668900582453, "einstein": 0.17220471693143607, "hypatia": 0.011709092132953687, "ibn_sina": 0.02172903688330456, "sagan": 0.765968549482974, "socrates": 0.013145935668749116}, {"ibn_sina": 0.02126161695214706, "hypatia": 0.01207962354733317, "sagan": 0.797890293724785, "socrates": 0.012675222197331353, "darwin": 0.016020162535730582, "einstein": 0.14007308104267285}, {"hypatia": 0.007936243487504838, "darwin": 0.01135349770151888, "sagan": 0.6936434850935634, "ibn_sina": 0.020539941452063656, "einstein": 0.2536600348986695, "socrates": 0.012866797366679675}, {"ibn_sina": 0.019529039819272027, "einstein": 0.2435322285063486, "darwin": 0.015317797755303637, "sagan": 0.7014638408768727, "hypatia": 0.009720230245967191, "socrates": 0.01043686279623587}, {"einstein": 0.2776649866539087, "sagan": 0.6734435472876866, "socrates": 0.014228178039756817, "ibn_sina": 0.016791583140019786, "darwin": 0.010458466972952583, "hypatia": 0.007413237905675454}, {"socrates": 0.011979513724613693, "ibn_sina": 0.021833827706045538, "hypatia": 0.012158760421589424, "einstein": 0.1587466447773045, "sagan": 0.7798897435614969, "darwin": 0.015391509808949941}], "prompt_tokens": 703, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.015242668900582453, "einstein": 0.17220471693143607, "hypatia": 0.011709092132953687, "ibn_sina": 0.02172903688330456, "sagan": 0.765968549482974, "socrates": 0.013145935668749116}, "confidence": 0.540852030151313}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.02126161695214706, "hypatia": 0.01207962354733317, "sagan": 0.797890293724785, "socrates": 0.012675222197331353, "darwin": 0.016020162535730582, "einstein": 0.14007308104267285}, "confidence": 0.5887176032546787}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.007936243487504838, "darwin": 0.01135349770151888, "sagan": 0.6936434850935634, "ibn_sina": 0.020539941452063656, "einstein": 0.2536600348986695, "socrates": 0.012866797366679675}, "confidence": 0.4555168327091435}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.019529039819272027, "einstein": 0.2435322285063486, "darwin": 0.015317797755303637, "sagan": 0.7014638408768727, "hypatia": 0.009720230245967191, "socrates": 0.01043686279623587}, "confidence": 0.4626146748212803}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.2776649866539087, "sagan": 0.6734435472876866, "socrates": 0.014228178039756817, "ibn_sina": 0.016791583140019786, "darwin": 0.010458466972952583, "hypatia": 0.007413237905675454}, "confidence": 0.4375273481671451}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.011979513724613693, "ibn_sina": 0.021833827706045538, "hypatia": 0.012158760421589424, "einstein": 0.1587466447773045, "sagan": 0.7798897435614969, "darwin": 0.015391509808949941}, "confidence": 0.5613201617812399}}, "usage": {"input_tokens": 703, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:05Z"}
{"id": "c-sagan-01-main-03", "uid": "c084", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "prompt_sha256_by_order": ["6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726"], "choice": "sagan", "probs": {"darwin": 0.0453812807186695, "einstein": 0.04434102912533597, "hypatia": 0.03770919577220205, "ibn_sina": 0.02366000954706661, "sagan": 0.8042562455398528, "socrates": 0.044652239296872995}, "confidence": 0.5857951575582392, "seconds": 0.25474403666642803, "seconds_all_orders": 1.5284642199985683, "seconds_by_order": [0.2544462059995567, 0.25523149999935413, 0.2550838009992731, 0.25405555900033505, 0.25424782700065407, 0.25539932699939527], "argmax_stable": true, "spread": {"darwin": 0.017837028533707437, "einstein": 0.03298463178628086, "hypatia": 0.011965079855020003, "ibn_sina": 0.00763041489288014, "sagan": 0.08381053470542765, "socrates": 0.03738990105060825}, "spread_2dp": {"darwin": 0.020000000000000004, "einstein": 0.030000000000000002, "hypatia": 0.020000000000000004, "ibn_sina": 0.0, "sagan": 0.07999999999999996, "socrates": 0.04}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0453812807186695, "einstein": 0.04434102912533597, "hypatia": 0.03770919577220205, "ibn_sina": 0.02366000954706661, "sagan": 0.8042562455398528, "socrates": 0.044652239296872995}, {"ibn_sina": 0.01602959465418647, "hypatia": 0.041615001434955695, "sagan": 0.8088103495665817, "socrates": 0.05131500824730725, "darwin": 0.05190545661944878, "einstein": 0.03032458947752005}, {"hypatia": 0.04673801994193233, "darwin": 0.04064151159329764, "sagan": 0.7588710790153106, "ibn_sina": 0.022238516967173917, "einstein": 0.0546288027103977, "socrates": 0.07688206977188786}, {"ibn_sina": 0.02019373771875576, "einstein": 0.04848913403978478, "darwin": 0.03834467053568272, "sagan": 0.7959590025768007, "hypatia": 0.03650866147751901, "socrates": 0.060504793651456876}, {"einstein": 0.021644170924116843, "sagan": 0.8426816137207382, "socrates": 0.03949216872127961, "ibn_sina": 0.019113334547109374, "darwin": 0.03406842808574134, "hypatia": 0.0430002840010145}, {"socrates": 0.05481617351744803, "ibn_sina": 0.01838924382776668, "hypatia": 0.03477294008691233, "einstein": 0.038801856598473516, "sagan": 0.8089054684622878, "darwin": 0.044314317507111636}], "prompt_tokens": 701, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.0453812807186695, "einstein": 0.04434102912533597, "hypatia": 0.03770919577220205, "ibn_sina": 0.02366000954706661, "sagan": 0.8042562455398528, "socrates": 0.044652239296872995}, "confidence": 0.5857951575582392}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.01602959465418647, "hypatia": 0.041615001434955695, "sagan": 0.8088103495665817, "socrates": 0.05131500824730725, "darwin": 0.05190545661944878, "einstein": 0.03032458947752005}, "confidence": 0.5948919100485999}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.04673801994193233, "darwin": 0.04064151159329764, "sagan": 0.7588710790153106, "ibn_sina": 0.022238516967173917, "einstein": 0.0546288027103977, "socrates": 0.07688206977188786}, "confidence": 0.5069334398961072}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.02019373771875576, "einstein": 0.04848913403978478, "darwin": 0.03834467053568272, "sagan": 0.7959590025768007, "hypatia": 0.03650866147751901, "socrates": 0.060504793651456876}, "confidence": 0.5713284917466558}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.021644170924116843, "sagan": 0.8426816137207382, "socrates": 0.03949216872127961, "ibn_sina": 0.019113334547109374, "darwin": 0.03406842808574134, "hypatia": 0.0430002840010145}, "confidence": 0.6586184864826411}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.05481617351744803, "ibn_sina": 0.01838924382776668, "hypatia": 0.03477294008691233, "einstein": 0.038801856598473516, "sagan": 0.8089054684622878, "darwin": 0.044314317507111636}, "confidence": 0.5948194410996915}}, "usage": {"input_tokens": 701, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:06Z"}
{"id": "c-sagan-01-starter-01", "uid": "c085", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "prompt_sha256_by_order": ["895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655"], "choice": "sagan", "probs": {"darwin": 0.009635499486917803, "einstein": 0.02384998536670939, "hypatia": 0.01183976651861703, "ibn_sina": 0.017801512020914956, "sagan": 0.9190951711347931, "socrates": 0.017778065472047706}, "confidence": 0.8154048781228287, "seconds": 0.19098685983287092, "seconds_all_orders": 1.1459211589972256, "seconds_by_order": [0.19115073699867935, 0.19062716099870158, 0.19040622300053656, 0.1912194560009084, 0.19145814399962546, 0.19105943799877423], "argmax_stable": true, "spread": {"darwin": 0.005364921584868656, "einstein": 0.016734519331075812, "hypatia": 0.002963361878151766, "ibn_sina": 0.009910763098037892, "sagan": 0.04714924199466952, "socrates": 0.016697152838521787}, "spread_2dp": {"darwin": 0.0, "einstein": 0.019999999999999997, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.039999999999999925, "socrates": 0.019999999999999997}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.009635499486917803, "einstein": 0.02384998536670939, "hypatia": 0.01183976651861703, "ibn_sina": 0.017801512020914956, "sagan": 0.9190951711347931, "socrates": 0.017778065472047706}, {"ibn_sina": 0.007890748922877064, "hypatia": 0.009837831734894617, "sagan": 0.9537136506592765, "socrates": 0.011955869861854599, "darwin": 0.007105930903030536, "einstein": 0.00949596791806669}, {"hypatia": 0.01186364572380542, "darwin": 0.011458859978380405, "sagan": 0.906564408664607, "ibn_sina": 0.015781940808389863, "einstein": 0.026230487249142502, "socrates": 0.028100657575674794}, {"ibn_sina": 0.008354498862620265, "einstein": 0.013666683026689509, "darwin": 0.0114146090736784, "sagan": 0.9451324335945683, "hypatia": 0.010028270705290457, "socrates": 0.011403504737153005}, {"einstein": 0.012719218471927819, "sagan": 0.9454657138845103, "socrates": 0.015937000456757083, "ibn_sina": 0.008621867369429671, "darwin": 0.006608494153978107, "hypatia": 0.01064770566339703}, {"socrates": 0.012649604817735458, "ibn_sina": 0.008176472457026657, "hypatia": 0.008900283845653653, "einstein": 0.018190311267397945, "sagan": 0.9401099118733394, "darwin": 0.011973415738846763}], "prompt_tokens": 473, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.009635499486917803, "einstein": 0.02384998536670939, "hypatia": 0.01183976651861703, "ibn_sina": 0.017801512020914956, "sagan": 0.9190951711347931, "socrates": 0.017778065472047706}, "confidence": 0.8154048781228287}}, "usage": {"input_tokens": 473, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.007890748922877064, "hypatia": 0.009837831734894617, "sagan": 0.9537136506592765, "socrates": 0.011955869861854599, "darwin": 0.007105930903030536, "einstein": 0.00949596791806669}, "confidence": 0.8920148617486067}}, "usage": {"input_tokens": 473, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.01186364572380542, "darwin": 0.011458859978380405, "sagan": 0.906564408664607, "ibn_sina": 0.015781940808389863, "einstein": 0.026230487249142502, "socrates": 0.028100657575674794}, "confidence": 0.788629400430958}}, "usage": {"input_tokens": 473, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.008354498862620265, "einstein": 0.013666683026689509, "darwin": 0.0114146090736784, "sagan": 0.9451324335945683, "hypatia": 0.010028270705290457, "socrates": 0.011403504737153005}, "confidence": 0.8726713508110269}}, "usage": {"input_tokens": 473, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.012719218471927819, "sagan": 0.9454657138845103, "socrates": 0.015937000456757083, "ibn_sina": 0.008621867369429671, "darwin": 0.006608494153978107, "hypatia": 0.01064770566339703}, "confidence": 0.8734630780732665}}, "usage": {"input_tokens": 473, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.012649604817735458, "ibn_sina": 0.008176472457026657, "hypatia": 0.008900283845653653, "einstein": 0.018190311267397945, "sagan": 0.9401099118733394, "darwin": 0.011973415738846763}, "confidence": 0.861504374520852}}, "usage": {"input_tokens": 473, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:08Z"}
{"id": "c-sagan-01-starter-02", "uid": "c086", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "prompt_sha256_by_order": ["30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263"], "choice": "sagan", "probs": {"darwin": 0.018941579562834505, "einstein": 0.08012163255103165, "hypatia": 0.011164517244341718, "ibn_sina": 0.012946576834420558, "sagan": 0.8604965968068767, "socrates": 0.016329097000494786}, "confidence": 0.6973498627132896, "seconds": 0.255635037500421, "seconds_all_orders": 1.5338102250025258, "seconds_by_order": [0.2558178350009257, 0.25557303599998704, 0.2565466979995108, 0.255488017000971, 0.25523901900123747, 0.2551456199998938], "argmax_stable": true, "spread": {"darwin": 0.010413284394613406, "einstein": 0.033955155381055785, "hypatia": 0.005342673640184975, "ibn_sina": 0.005705581965151982, "sagan": 0.04923344252993345, "socrates": 0.0048892133521380755}, "spread_2dp": {"darwin": 0.01, "einstein": 0.03, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.050000000000000044, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.018941579562834505, "einstein": 0.08012163255103165, "hypatia": 0.011164517244341718, "ibn_sina": 0.012946576834420558, "sagan": 0.8604965968068767, "socrates": 0.016329097000494786}, {"ibn_sina": 0.00877448569792259, "hypatia": 0.00993963653249538, "sagan": 0.8493338017531273, "socrates": 0.013467927792481692, "darwin": 0.01623510379449379, "einstein": 0.10224904442947921}, {"hypatia": 0.008444218944397836, "darwin": 0.013727341005965513, "sagan": 0.85299491866784, "ibn_sina": 0.011895282466965357, "einstein": 0.0980865839173217, "socrates": 0.014851654997509471}, {"ibn_sina": 0.011213772521070007, "einstein": 0.08278329971695139, "darwin": 0.0174235830890423, "sagan": 0.865383047694119, "hypatia": 0.010110095024760186, "socrates": 0.013086201954057101}, {"einstein": 0.06829388904842343, "sagan": 0.8985672442830608, "socrates": 0.01143988364835671, "ibn_sina": 0.007240994869268576, "darwin": 0.008528295168221098, "hypatia": 0.005929692982669274}, {"socrates": 0.012637280949102847, "ibn_sina": 0.008017860030210358, "hypatia": 0.01127236662285425, "einstein": 0.06885588276637583, "sagan": 0.882985900845112, "darwin": 0.016230708786344732}], "prompt_tokens": 713, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.018941579562834505, "einstein": 0.08012163255103165, "hypatia": 0.011164517244341718, "ibn_sina": 0.012946576834420558, "sagan": 0.8604965968068767, "socrates": 0.016329097000494786}, "confidence": 0.6973498627132896}}, "usage": {"input_tokens": 713, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.00877448569792259, "hypatia": 0.00993963653249538, "sagan": 0.8493338017531273, "socrates": 0.013467927792481692, "darwin": 0.01623510379449379, "einstein": 0.10224904442947921}, "confidence": 0.6789322306420779}}, "usage": {"input_tokens": 713, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.008444218944397836, "darwin": 0.013727341005965513, "sagan": 0.85299491866784, "ibn_sina": 0.011895282466965357, "einstein": 0.0980865839173217, "socrates": 0.014851654997509471}, "confidence": 0.6854117480122405}}, "usage": {"input_tokens": 713, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.011213772521070007, "einstein": 0.08278329971695139, "darwin": 0.0174235830890423, "sagan": 0.865383047694119, "hypatia": 0.010110095024760186, "socrates": 0.013086201954057101}, "confidence": 0.7077324239118346}}, "usage": {"input_tokens": 713, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.06829388904842343, "sagan": 0.8985672442830608, "socrates": 0.01143988364835671, "ibn_sina": 0.007240994869268576, "darwin": 0.008528295168221098, "hypatia": 0.005929692982669274}, "confidence": 0.7748540125620902}}, "usage": {"input_tokens": 713, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.012637280949102847, "ibn_sina": 0.008017860030210358, "hypatia": 0.01127236662285425, "einstein": 0.06885588276637583, "sagan": 0.882985900845112, "darwin": 0.016230708786344732}, "confidence": 0.7420236673468343}}, "usage": {"input_tokens": 713, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:09Z"}
{"id": "c-sagan-01-starter-02", "uid": "c087", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "prompt_sha256_by_order": ["dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a"], "choice": "sagan", "probs": {"darwin": 0.008740522136806423, "einstein": 0.048854455468735354, "hypatia": 0.013081971040846654, "ibn_sina": 0.007105282224187149, "sagan": 0.9102774106775993, "socrates": 0.01194035845182518}, "confidence": 0.7977187769176508, "seconds": 0.19304743766709484, "seconds_all_orders": 1.158284626002569, "seconds_by_order": [0.1937982550007291, 0.19285736299934797, 0.19304834100148582, 0.19236251700021967, 0.19286260199987737, 0.19335554800090904], "argmax_stable": true, "spread": {"darwin": 0.005073159187027306, "einstein": 0.22283455248201084, "hypatia": 0.020435138200997207, "ibn_sina": 0.004851941654583041, "sagan": 0.22178785001985313, "socrates": 0.006387441238161347}, "spread_2dp": {"darwin": 0.0, "einstein": 0.22000000000000003, "hypatia": 0.019999999999999997, "ibn_sina": 0.01, "sagan": 0.22000000000000008, "socrates": 0.0}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.008740522136806423, "einstein": 0.048854455468735354, "hypatia": 0.013081971040846654, "ibn_sina": 0.007105282224187149, "sagan": 0.9102774106775993, "socrates": 0.01194035845182518}, {"ibn_sina": 0.007729059463859636, "hypatia": 0.017013347185794857, "sagan": 0.6884895606577461, "socrates": 0.008566446737118335, "darwin": 0.006512578004734869, "einstein": 0.2716890079507462}, {"hypatia": 0.022186107878275546, "darwin": 0.008282625393084813, "sagan": 0.8596274243775628, "ibn_sina": 0.00790551008717153, "einstein": 0.08959760275236031, "socrates": 0.012400729511545014}, {"ibn_sina": 0.00978257492113048, "einstein": 0.09062778827322035, "darwin": 0.0094355962786388, "sagan": 0.8643047945748731, "hypatia": 0.014414176498554247, "socrates": 0.011435069453583021}, {"einstein": 0.053898566195351494, "sagan": 0.887618654866583, "socrates": 0.014953887975279682, "ibn_sina": 0.004930633266547439, "darwin": 0.005081148454394474, "hypatia": 0.03351710924184386}, {"socrates": 0.009078144076355113, "ibn_sina": 0.006757421392018261, "hypatia": 0.02414019224517735, "einstein": 0.04960813761500568, "sagan": 0.9002617970300217, "darwin": 0.01015430764142178}], "prompt_tokens": 493, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.008740522136806423, "einstein": 0.048854455468735354, "hypatia": 0.013081971040846654, "ibn_sina": 0.007105282224187149, "sagan": 0.9102774106775993, "socrates": 0.01194035845182518}, "confidence": 0.7977187769176508}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.007729059463859636, "hypatia": 0.017013347185794857, "sagan": 0.6884895606577461, "socrates": 0.008566446737118335, "darwin": 0.006512578004734869, "einstein": 0.2716890079507462}, "confidence": 0.45795733864070176}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.022186107878275546, "darwin": 0.008282625393084813, "sagan": 0.8596274243775628, "ibn_sina": 0.00790551008717153, "einstein": 0.08959760275236031, "socrates": 0.012400729511545014}, "confidence": 0.6973169675311812}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.00978257492113048, "einstein": 0.09062778827322035, "darwin": 0.0094355962786388, "sagan": 0.8643047945748731, "hypatia": 0.014414176498554247, "socrates": 0.011435069453583021}, "confidence": 0.7069113189750038}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.053898566195351494, "sagan": 0.887618654866583, "socrates": 0.014953887975279682, "ibn_sina": 0.004930633266547439, "darwin": 0.005081148454394474, "hypatia": 0.03351710924184386}, "confidence": 0.7506028917959302}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.009078144076355113, "ibn_sina": 0.006757421392018261, "hypatia": 0.02414019224517735, "einstein": 0.04960813761500568, "sagan": 0.9002617970300217, "darwin": 0.01015430764142178}, "confidence": 0.7764954457580853}}, "usage": {"input_tokens": 493, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:10Z"}
{"id": "c-sagan-01-starter-03", "uid": "c088", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "prompt_sha256_by_order": ["ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f"], "choice": "darwin", "probs": {"darwin": 0.844539123431534, "einstein": 0.023704531279988037, "hypatia": 0.006371411281665645, "ibn_sina": 0.02372225509101266, "sagan": 0.09293018143161128, "socrates": 0.008732497484188386}, "confidence": 0.6677486214535054, "seconds": 0.1867141064997971, "seconds_all_orders": 1.1202846389987826, "seconds_by_order": [0.1867434730011155, 0.18701270999918052, 0.18630365699937101, 0.18641788599961728, 0.1873498679997283, 0.18645704499977], "argmax_stable": true, "spread": {"darwin": 0.08420983660478876, "einstein": 0.013522941581771718, "hypatia": 0.0034391069717576474, "ibn_sina": 0.016315447625726988, "sagan": 0.08597983467142746, "socrates": 0.003263224360868281}, "spread_2dp": {"darwin": 0.08999999999999997, "einstein": 0.019999999999999997, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.08000000000000002, "socrates": 0.0}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.844539123431534, "einstein": 0.023704531279988037, "hypatia": 0.006371411281665645, "ibn_sina": 0.02372225509101266, "sagan": 0.09293018143161128, "socrates": 0.008732497484188386}, {"ibn_sina": 0.0078982791491913, "hypatia": 0.007066843204900388, "sagan": 0.05649941533311813, "socrates": 0.008006252062072897, "darwin": 0.8959388730037816, "einstein": 0.024590337246935635}, {"hypatia": 0.00820468169694475, "darwin": 0.8958358318094684, "sagan": 0.067067969768182, "ibn_sina": 0.008592684295498905, "einstein": 0.013892284092396186, "socrates": 0.006406548337509741}, {"ibn_sina": 0.009665925509671179, "einstein": 0.020839625201180115, "darwin": 0.8117290363989929, "sagan": 0.1424792500045456, "hypatia": 0.0074585464224358545, "socrates": 0.007827616463174326}, {"einstein": 0.027415225674167903, "sagan": 0.07651291211599051, "socrates": 0.005469273123320105, "ibn_sina": 0.008320417445664266, "darwin": 0.8743147692161055, "hypatia": 0.0079674024247516}, {"socrates": 0.006343651550226979, "ibn_sina": 0.0074068074652856725, "hypatia": 0.004765574725187103, "einstein": 0.015925900017238547, "sagan": 0.1234320423883311, "darwin": 0.8421260238537306}], "prompt_tokens": 511, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"darwin": 0.844539123431534, "einstein": 0.023704531279988037, "hypatia": 0.006371411281665645, "ibn_sina": 0.02372225509101266, "sagan": 0.09293018143161128, "socrates": 0.008732497484188386}, "confidence": 0.6677486214535054}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.0078982791491913, "hypatia": 0.007066843204900388, "sagan": 0.05649941533311813, "socrates": 0.008006252062072897, "darwin": 0.8959388730037816, "einstein": 0.024590337246935635}, "confidence": 0.7680157071240054}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.00820468169694475, "darwin": 0.8958358318094684, "sagan": 0.067067969768182, "ibn_sina": 0.008592684295498905, "einstein": 0.013892284092396186, "socrates": 0.006406548337509741}, "confidence": 0.7688741686799714}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"ibn_sina": 0.009665925509671179, "einstein": 0.020839625201180115, "darwin": 0.8117290363989929, "sagan": 0.1424792500045456, "hypatia": 0.0074585464224358545, "socrates": 0.007827616463174326}, "confidence": 0.6158187841645717}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"einstein": 0.027415225674167903, "sagan": 0.07651291211599051, "socrates": 0.005469273123320105, "ibn_sina": 0.008320417445664266, "darwin": 0.8743147692161055, "hypatia": 0.0079674024247516}, "confidence": 0.7254337093420148}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.006343651550226979, "ibn_sina": 0.0074068074652856725, "hypatia": 0.004765574725187103, "einstein": 0.015925900017238547, "sagan": 0.1234320423883311, "darwin": 0.8421260238537306}, "confidence": 0.6697397882145852}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:11Z"}
{"id": "c-sagan-01-starter-03", "uid": "c089", "task": "kperm", "label": "sagan", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "prompt_sha256_by_order": ["d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f"], "choice": "einstein", "probs": {"darwin": 0.06859212191312497, "einstein": 0.8413718748924817, "hypatia": 0.006950041213773582, "ibn_sina": 0.009638220300114483, "sagan": 0.051812523362848764, "socrates": 0.021635218317656465}, "confidence": 0.659086395593128, "seconds": 0.2242011565000818, "seconds_all_orders": 1.3452069390004908, "seconds_by_order": [0.22468164099882415, 0.2235836300005758, 0.22416491500007396, 0.22441685299963865, 0.22441268300099182, 0.22394721700038644], "argmax_stable": true, "spread": {"darwin": 0.07245129089996405, "einstein": 0.12709069631639713, "hypatia": 0.005759415187504576, "ibn_sina": 0.007252673424303557, "sagan": 0.06978898211524817, "socrates": 0.023127188540520822}, "spread_2dp": {"darwin": 0.07, "einstein": 0.12, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.06999999999999999, "socrates": 0.02}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.06859212191312497, "einstein": 0.8413718748924817, "hypatia": 0.006950041213773582, "ibn_sina": 0.009638220300114483, "sagan": 0.051812523362848764, "socrates": 0.021635218317656465}, {"ibn_sina": 0.012671845220176831, "hypatia": 0.010145467418990396, "sagan": 0.05048034732409577, "socrates": 0.018771321965185887, "darwin": 0.05428718417629666, "einstein": 0.8536438338952544}, {"hypatia": 0.007444604250849489, "darwin": 0.03881093669430522, "sagan": 0.04556100765845299, "ibn_sina": 0.01036693455644406, "einstein": 0.8659144682825235, "socrates": 0.031902048557424775}, {"ibn_sina": 0.014558860612023524, "einstein": 0.8328399836826776, "darwin": 0.03163253342034988, "sagan": 0.06912974492708676, "hypatia": 0.011550416993575204, "socrates": 0.04028846036428707}, {"einstein": 0.8731341418965144, "sagan": 0.07089243784310291, "socrates": 0.01716127182376625, "ibn_sina": 0.007306187187719967, "darwin": 0.025714959442825876, "hypatia": 0.005791001806070628}, {"socrates": 0.02161269720011607, "ibn_sina": 0.011588897569991078, "hypatia": 0.007238719533284525, "einstein": 0.7460434455801173, "sagan": 0.11534998977370116, "darwin": 0.09816625034278992}], "prompt_tokens": 615, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"darwin": 0.06859212191312497, "einstein": 0.8413718748924817, "hypatia": 0.006950041213773582, "ibn_sina": 0.009638220300114483, "sagan": 0.051812523362848764, "socrates": 0.021635218317656465}, "confidence": 0.659086395593128}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.012671845220176831, "hypatia": 0.010145467418990396, "sagan": 0.05048034732409577, "socrates": 0.018771321965185887, "darwin": 0.05428718417629666, "einstein": 0.8536438338952544}, "confidence": 0.681782833213475}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"hypatia": 0.007444604250849489, "darwin": 0.03881093669430522, "sagan": 0.04556100765845299, "ibn_sina": 0.01036693455644406, "einstein": 0.8659144682825235, "socrates": 0.031902048557424775}, "confidence": 0.7054847161283895}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"ibn_sina": 0.014558860612023524, "einstein": 0.8328399836826776, "darwin": 0.03163253342034988, "sagan": 0.06912974492708676, "hypatia": 0.011550416993575204, "socrates": 0.04028846036428707}, "confidence": 0.6416446117820884}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"einstein": 0.8731341418965144, "sagan": 0.07089243784310291, "socrates": 0.01716127182376625, "ibn_sina": 0.007306187187719967, "darwin": 0.025714959442825876, "hypatia": 0.005791001806070628}, "confidence": 0.7221179823419153}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "einstein", "probabilities": {"socrates": 0.02161269720011607, "ibn_sina": 0.011588897569991078, "hypatia": 0.007238719533284525, "einstein": 0.7460434455801173, "sagan": 0.11534998977370116, "darwin": 0.09816625034278992}, "confidence": 0.49621223899341144}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:13Z"}
{"id": "c-socrates-01-dessert-01", "uid": "c090", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "prompt_sha256_by_order": ["b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee"], "choice": "socrates", "probs": {"darwin": 0.019640083304400383, "einstein": 0.028120260838927713, "hypatia": 0.040499986592254504, "ibn_sina": 0.10079425766633998, "sagan": 0.02943980425351498, "socrates": 0.7815056073445624}, "confidence": 0.5495127155437759, "seconds": 0.2567475968335202, "seconds_all_orders": 1.5404855810011213, "seconds_by_order": [0.25642975899972953, 0.2572150230007537, 0.2567857770009141, 0.2564777989991853, 0.2562184310008888, 0.25735879199964984], "argmax_stable": true, "spread": {"darwin": 0.007192984203998103, "einstein": 0.018002347034915867, "hypatia": 0.018647742533794108, "ibn_sina": 0.05173836067112211, "sagan": 0.020620403266818882, "socrates": 0.07755646496209789}, "spread_2dp": {"darwin": 0.01, "einstein": 0.009999999999999998, "hypatia": 0.020000000000000004, "ibn_sina": 0.06, "sagan": 0.02, "socrates": 0.07999999999999996}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.019640083304400383, "einstein": 0.028120260838927713, "hypatia": 0.040499986592254504, "ibn_sina": 0.10079425766633998, "sagan": 0.02943980425351498, "socrates": 0.7815056073445624}, {"ibn_sina": 0.07453236862934687, "hypatia": 0.027271864765527745, "sagan": 0.020508152654903467, "socrates": 0.8359373244055369, "darwin": 0.019220479181075143, "einstein": 0.022529810363609848}, {"hypatia": 0.0329059107743324, "darwin": 0.01514260781315101, "sagan": 0.02064221017938594, "ibn_sina": 0.12627072930046898, "einstein": 0.024227848412467664, "socrates": 0.7808106935201939}, {"ibn_sina": 0.08578156001759954, "einstein": 0.02608434321229317, "darwin": 0.020881106116098055, "sagan": 0.025199909722985712, "hypatia": 0.04009313141843085, "socrates": 0.8019599495125926}, {"einstein": 0.03306716506524256, "sagan": 0.04112855592172235, "socrates": 0.758380859443439, "ibn_sina": 0.10752065203422208, "darwin": 0.014084113886017205, "hypatia": 0.045818653649356736}, {"socrates": 0.79874000503132, "ibn_sina": 0.10598480741924723, "hypatia": 0.04591960729932185, "einstein": 0.01506481803032669, "sagan": 0.020602640307684227, "darwin": 0.013688121912099952}], "prompt_tokens": 723, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.019640083304400383, "einstein": 0.028120260838927713, "hypatia": 0.040499986592254504, "ibn_sina": 0.10079425766633998, "sagan": 0.02943980425351498, "socrates": 0.7815056073445624}, "confidence": 0.5495127155437759}}, "usage": {"input_tokens": 723, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.07453236862934687, "hypatia": 0.027271864765527745, "sagan": 0.020508152654903467, "socrates": 0.8359373244055369, "darwin": 0.019220479181075143, "einstein": 0.022529810363609848}, "confidence": 0.6476651708990382}}, "usage": {"input_tokens": 723, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.0329059107743324, "darwin": 0.01514260781315101, "sagan": 0.02064221017938594, "ibn_sina": 0.12627072930046898, "einstein": 0.024227848412467664, "socrates": 0.7808106935201939}, "confidence": 0.5535217878503407}}, "usage": {"input_tokens": 723, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.08578156001759954, "einstein": 0.02608434321229317, "darwin": 0.020881106116098055, "sagan": 0.025199909722985712, "hypatia": 0.04009313141843085, "socrates": 0.8019599495125926}, "confidence": 0.5846285738220685}}, "usage": {"input_tokens": 723, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.03306716506524256, "sagan": 0.04112855592172235, "socrates": 0.758380859443439, "ibn_sina": 0.10752065203422208, "darwin": 0.014084113886017205, "hypatia": 0.045818653649356736}, "confidence": 0.5101419104654629}}, "usage": {"input_tokens": 723, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.79874000503132, "ibn_sina": 0.10598480741924723, "hypatia": 0.04591960729932185, "einstein": 0.01506481803032669, "sagan": 0.020602640307684227, "darwin": 0.013688121912099952}, "confidence": 0.5825989211044806}}, "usage": {"input_tokens": 723, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:14Z"}
{"id": "c-socrates-01-dessert-01", "uid": "c091", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "prompt_sha256_by_order": ["42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512"], "choice": "socrates", "probs": {"darwin": 0.016540322430792707, "einstein": 0.016707956659196808, "hypatia": 0.017303598555222264, "ibn_sina": 0.016707956659196808, "sagan": 0.03425324384591989, "socrates": 0.8984869218496716}, "confidence": 0.7715000102432704, "seconds": 0.1926683606664786, "seconds_all_orders": 1.1560101639988716, "seconds_by_order": [0.1929146820002643, 0.19207542900039698, 0.1928253030000633, 0.19239626599846815, 0.19301626100059366, 0.1927822229990852], "argmax_stable": true, "spread": {"darwin": 0.007559213952758666, "einstein": 0.007851095647027524, "hypatia": 0.004128088104966647, "ibn_sina": 0.012225990532791508, "sagan": 0.018581235281934824, "socrates": 0.035572381288077715}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.009999999999999998, "socrates": 0.030000000000000027}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.016540322430792707, "einstein": 0.016707956659196808, "hypatia": 0.017303598555222264, "ibn_sina": 0.016707956659196808, "sagan": 0.03425324384591989, "socrates": 0.8984869218496716}, {"ibn_sina": 0.02232557028564679, "hypatia": 0.014762274786125984, "sagan": 0.022289708875841044, "socrates": 0.9143986072553321, "darwin": 0.01230041647851305, "einstein": 0.013923422318541018}, {"hypatia": 0.01889036289109263, "darwin": 0.008981108478034041, "sagan": 0.015672008563985068, "ibn_sina": 0.010099579752855282, "einstein": 0.01229763717628367, "socrates": 0.9340593031377493}, {"ibn_sina": 0.01143495167966875, "einstein": 0.008856861012169284, "darwin": 0.009245472222434641, "sagan": 0.022881844741269197, "hypatia": 0.015246268709432177, "socrates": 0.932334601635026}, {"einstein": 0.01455861217503784, "sagan": 0.01881766357981144, "socrates": 0.9228727558542159, "ibn_sina": 0.01584674460545019, "darwin": 0.009927443042837273, "hypatia": 0.017976780742647303}, {"socrates": 0.9176419266087952, "ibn_sina": 0.014891512974772717, "hypatia": 0.017115276232117983, "einstein": 0.011719049982051824, "sagan": 0.022971393608180485, "darwin": 0.01566084059408181}], "prompt_tokens": 489, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.016540322430792707, "einstein": 0.016707956659196808, "hypatia": 0.017303598555222264, "ibn_sina": 0.016707956659196808, "sagan": 0.03425324384591989, "socrates": 0.8984869218496716}, "confidence": 0.7715000102432704}}, "usage": {"input_tokens": 489, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.02232557028564679, "hypatia": 0.014762274786125984, "sagan": 0.022289708875841044, "socrates": 0.9143986072553321, "darwin": 0.01230041647851305, "einstein": 0.013923422318541018}, "confidence": 0.8052197942227679}}, "usage": {"input_tokens": 489, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.01889036289109263, "darwin": 0.008981108478034041, "sagan": 0.015672008563985068, "ibn_sina": 0.010099579752855282, "einstein": 0.01229763717628367, "socrates": 0.9340593031377493}, "confidence": 0.8480837597698752}}, "usage": {"input_tokens": 489, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.01143495167966875, "einstein": 0.008856861012169284, "darwin": 0.009245472222434641, "sagan": 0.022881844741269197, "hypatia": 0.015246268709432177, "socrates": 0.932334601635026}, "confidence": 0.844358221357345}}, "usage": {"input_tokens": 489, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.01455861217503784, "sagan": 0.01881766357981144, "socrates": 0.9228727558542159, "ibn_sina": 0.01584674460545019, "darwin": 0.009927443042837273, "hypatia": 0.017976780742647303}, "confidence": 0.8235196230818197}}, "usage": {"input_tokens": 489, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.9176419266087952, "ibn_sina": 0.014891512974772717, "hypatia": 0.017115276232117983, "einstein": 0.011719049982051824, "sagan": 0.022971393608180485, "darwin": 0.01566084059408181}, "confidence": 0.8121900139532535}}, "usage": {"input_tokens": 489, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:15Z"}
{"id": "c-socrates-01-dessert-01", "uid": "c092", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "prompt_sha256_by_order": ["67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261"], "choice": "socrates", "probs": {"darwin": 0.2594804436815537, "einstein": 0.09210771650140345, "hypatia": 0.0769354027651901, "ibn_sina": 0.13060310358265712, "sagan": 0.10892670991133646, "socrates": 0.3319466235578591}, "confidence": 0.06501249757589753, "seconds": 0.22513272533372705, "seconds_all_orders": 1.3507963520023623, "seconds_by_order": [0.22508824700162222, 0.2253040659998078, 0.2251406670002325, 0.2259158100005152, 0.22465001099953952, 0.22469755100064503], "argmax_stable": false, "spread": {"darwin": 0.10002420581931676, "einstein": 0.04048780941352487, "hypatia": 0.011870228498251464, "ibn_sina": 0.13176257258181684, "sagan": 0.06962533511893922, "socrates": 0.10664223670649114}, "spread_2dp": {"darwin": 0.1, "einstein": 0.04000000000000001, "hypatia": 0.01999999999999999, "ibn_sina": 0.13, "sagan": 0.07, "socrates": 0.09999999999999998}, "choices_by_order": ["socrates", "socrates", "darwin", "socrates", "socrates", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.2594804436815537, "einstein": 0.09210771650140345, "hypatia": 0.0769354027651901, "ibn_sina": 0.13060310358265712, "sagan": 0.10892670991133646, "socrates": 0.3319466235578591}, {"ibn_sina": 0.2183907986974427, "hypatia": 0.07886193111788503, "sagan": 0.09439132936294972, "socrates": 0.3202961745575755, "darwin": 0.21664029805134483, "einstein": 0.07141946821280219}, {"hypatia": 0.07415293441436555, "darwin": 0.3166645038706616, "sagan": 0.1607443811581098, "ibn_sina": 0.0935809375453802, "einstein": 0.0787696666866513, "socrates": 0.27608757632483155}, {"ibn_sina": 0.08662822611562586, "einstein": 0.06253487752865154, "darwin": 0.2969486447969013, "sagan": 0.09111904603917058, "hypatia": 0.08003939248832803, "socrates": 0.3827298130313227}, {"einstein": 0.10302268694217641, "sagan": 0.11274448696566797, "socrates": 0.2984425733122148, "ibn_sina": 0.17523422384698598, "darwin": 0.2245328660203378, "hypatia": 0.08602316291261701}, {"socrates": 0.2945186849709583, "ibn_sina": 0.1264979843310422, "hypatia": 0.0747268226866029, "einstein": 0.08489372540300921, "sagan": 0.11086781066579458, "darwin": 0.3084949719425928}], "prompt_tokens": 625, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.2594804436815537, "einstein": 0.09210771650140345, "hypatia": 0.0769354027651901, "ibn_sina": 0.13060310358265712, "sagan": 0.10892670991133646, "socrates": 0.3319466235578591}, "confidence": 0.06501249757589753}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.2183907986974427, "hypatia": 0.07886193111788503, "sagan": 0.09439132936294972, "socrates": 0.3202961745575755, "darwin": 0.21664029805134483, "einstein": 0.07141946821280219}, "confidence": 0.06093624017205483}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"hypatia": 0.07415293441436555, "darwin": 0.3166645038706616, "sagan": 0.1607443811581098, "ibn_sina": 0.0935809375453802, "einstein": 0.0787696666866513, "socrates": 0.27608757632483155}, "confidence": 0.06736026859664235}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.08662822611562586, "einstein": 0.06253487752865154, "darwin": 0.2969486447969013, "sagan": 0.09111904603917058, "hypatia": 0.08003939248832803, "socrates": 0.3827298130313227}, "confidence": 0.11294158335779203}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.10302268694217641, "sagan": 0.11274448696566797, "socrates": 0.2984425733122148, "ibn_sina": 0.17523422384698598, "darwin": 0.2245328660203378, "hypatia": 0.08602316291261701}, "confidence": 0.04109798634304873}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "darwin", "probabilities": {"socrates": 0.2945186849709583, "ibn_sina": 0.1264979843310422, "hypatia": 0.0747268226866029, "einstein": 0.08489372540300921, "sagan": 0.11086781066579458, "darwin": 0.3084949719425928}, "confidence": 0.06759382916099596}}, "usage": {"input_tokens": 625, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:17Z"}
{"id": "c-socrates-01-dessert-02", "uid": "c093", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "prompt_sha256_by_order": ["14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e"], "choice": "sagan", "probs": {"darwin": 0.022720150424847433, "einstein": 0.12778786554480187, "hypatia": 0.028206631178176622, "ibn_sina": 0.17785960824649583, "sagan": 0.5864791811523556, "socrates": 0.05694656345332262}, "confidence": 0.2757716069417483, "seconds": 0.2221582648335243, "seconds_all_orders": 1.3329495890011458, "seconds_by_order": [0.2222419509998872, 0.22257679799986363, 0.22251763799977198, 0.2214593770004285, 0.2221165419996396, 0.2220372830015549], "argmax_stable": true, "spread": {"darwin": 0.009844590334801454, "einstein": 0.059557449350820024, "hypatia": 0.009990131040238916, "ibn_sina": 0.14698868438108964, "sagan": 0.16630411907813836, "socrates": 0.03368640666603821}, "spread_2dp": {"darwin": 0.01, "einstein": 0.06, "hypatia": 0.009999999999999998, "ibn_sina": 0.15000000000000002, "sagan": 0.16000000000000003, "socrates": 0.03}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.022720150424847433, "einstein": 0.12778786554480187, "hypatia": 0.028206631178176622, "ibn_sina": 0.17785960824649583, "sagan": 0.5864791811523556, "socrates": 0.05694656345332262}, {"ibn_sina": 0.20031180555384961, "hypatia": 0.018216500137937706, "sagan": 0.6001973525224005, "socrates": 0.04864447296673596, "darwin": 0.015041137922131543, "einstein": 0.11758873089694476}, {"hypatia": 0.0253009828614734, "darwin": 0.016019186908490572, "sagan": 0.6636807425918759, "ibn_sina": 0.12428189936291184, "einstein": 0.09903828651322835, "socrates": 0.07167890176201992}, {"ibn_sina": 0.16304263288425355, "einstein": 0.06823041619398185, "darwin": 0.01627554230847348, "sagan": 0.651296770875404, "hypatia": 0.020351167861433178, "socrates": 0.08080346987645398}, {"einstein": 0.10239792083071779, "sagan": 0.5871945097021181, "socrates": 0.08040697017122662, "ibn_sina": 0.19069334244553396, "darwin": 0.012875560090045979, "hypatia": 0.02643169676035763}, {"socrates": 0.08233087963277418, "ibn_sina": 0.2712705837440015, "hypatia": 0.027551631440451408, "einstein": 0.1032295045694423, "sagan": 0.49737662351373757, "darwin": 0.018240777099593035}], "prompt_tokens": 593, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"darwin": 0.022720150424847433, "einstein": 0.12778786554480187, "hypatia": 0.028206631178176622, "ibn_sina": 0.17785960824649583, "sagan": 0.5864791811523556, "socrates": 0.05694656345332262}, "confidence": 0.2757716069417483}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.20031180555384961, "hypatia": 0.018216500137937706, "sagan": 0.6001973525224005, "socrates": 0.04864447296673596, "darwin": 0.015041137922131543, "einstein": 0.11758873089694476}, "confidence": 0.3005357830126908}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"hypatia": 0.0253009828614734, "darwin": 0.016019186908490572, "sagan": 0.6636807425918759, "ibn_sina": 0.12428189936291184, "einstein": 0.09903828651322835, "socrates": 0.07167890176201992}, "confidence": 0.36611358379936976}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"ibn_sina": 0.16304263288425355, "einstein": 0.06823041619398185, "darwin": 0.01627554230847348, "sagan": 0.651296770875404, "hypatia": 0.020351167861433178, "socrates": 0.08080346987645398}, "confidence": 0.3551608451672754}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"einstein": 0.10239792083071779, "sagan": 0.5871945097021181, "socrates": 0.08040697017122662, "ibn_sina": 0.19069334244553396, "darwin": 0.012875560090045979, "hypatia": 0.02643169676035763}, "confidence": 0.27877164731343385}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "sagan", "probabilities": {"socrates": 0.08233087963277418, "ibn_sina": 0.2712705837440015, "hypatia": 0.027551631440451408, "einstein": 0.1032295045694423, "sagan": 0.49737662351373757, "darwin": 0.018240777099593035}, "confidence": 0.2073973095060687}}, "usage": {"input_tokens": 593, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:18Z"}
{"id": "c-socrates-01-dessert-02", "uid": "c094", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "prompt_sha256_by_order": ["d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6"], "choice": "socrates", "probs": {"darwin": 0.008310972630037829, "einstein": 0.007138921174963887, "hypatia": 0.007843739578326665, "ibn_sina": 0.0063168119707171415, "sagan": 0.009662320242392853, "socrates": 0.9607272344035616}, "confidence": 0.9079739706193937, "seconds": 0.22794568050024586, "seconds_all_orders": 1.367674083001475, "seconds_by_order": [0.22831217000020843, 0.2276566559994535, 0.22804838299998664, 0.22741554800086305, 0.22798941400105832, 0.22825191199990513], "argmax_stable": true, "spread": {"darwin": 0.002200118289153284, "einstein": 0.0021740182806331814, "hypatia": 0.005979929746023132, "ibn_sina": 0.002088527085910859, "sagan": 0.006674406871529762, "socrates": 0.017296152939362375}, "spread_2dp": {"darwin": 0.0, "einstein": 0.01, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.01, "socrates": 0.020000000000000018}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.008310972630037829, "einstein": 0.007138921174963887, "hypatia": 0.007843739578326665, "ibn_sina": 0.0063168119707171415, "sagan": 0.009662320242392853, "socrates": 0.9607272344035616}, {"ibn_sina": 0.007278121885534748, "hypatia": 0.010843250167945885, "sagan": 0.010437601654581262, "socrates": 0.9566710216403327, "darwin": 0.00823588570676557, "einstein": 0.006534118944839816}, {"hypatia": 0.01236159589195654, "darwin": 0.008057007182723076, "sagan": 0.008494043202393164, "ibn_sina": 0.0054545795771390095, "einstein": 0.006663353878394701, "socrates": 0.9589694202673935}, {"ibn_sina": 0.005411743594484972, "einstein": 0.004964902894330705, "darwin": 0.006110854340884545, "sagan": 0.008705304407319208, "hypatia": 0.009054204583731311, "socrates": 0.9657529901792493}, {"einstein": 0.006951261940436606, "sagan": 0.015168450073922927, "socrates": 0.9484568372398869, "ibn_sina": 0.007500270680395831, "darwin": 0.008099510741007893, "hypatia": 0.013823669324349797}, {"socrates": 0.9588617222134324, "ibn_sina": 0.0060653753154202555, "hypatia": 0.009207237282430619, "einstein": 0.0058031084897838, "sagan": 0.0117599545203942, "darwin": 0.008302602178538798}], "prompt_tokens": 653, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.008310972630037829, "einstein": 0.007138921174963887, "hypatia": 0.007843739578326665, "ibn_sina": 0.0063168119707171415, "sagan": 0.009662320242392853, "socrates": 0.9607272344035616}, "confidence": 0.9079739706193937}}, "usage": {"input_tokens": 653, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.007278121885534748, "hypatia": 0.010843250167945885, "sagan": 0.010437601654581262, "socrates": 0.9566710216403327, "darwin": 0.00823588570676557, "einstein": 0.006534118944839816}, "confidence": 0.8987313505969639}}, "usage": {"input_tokens": 653, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.01236159589195654, "darwin": 0.008057007182723076, "sagan": 0.008494043202393164, "ibn_sina": 0.0054545795771390095, "einstein": 0.006663353878394701, "socrates": 0.9589694202673935}, "confidence": 0.9039836499026986}}, "usage": {"input_tokens": 653, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.005411743594484972, "einstein": 0.004964902894330705, "darwin": 0.006110854340884545, "sagan": 0.008705304407319208, "hypatia": 0.009054204583731311, "socrates": 0.9657529901792493}, "confidence": 0.9195134545070651}}, "usage": {"input_tokens": 653, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.006951261940436606, "sagan": 0.015168450073922927, "socrates": 0.9484568372398869, "ibn_sina": 0.007500270680395831, "darwin": 0.008099510741007893, "hypatia": 0.013823669324349797}, "confidence": 0.8801940687944929}}, "usage": {"input_tokens": 653, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.9588617222134324, "ibn_sina": 0.0060653753154202555, "hypatia": 0.009207237282430619, "einstein": 0.0058031084897838, "sagan": 0.0117599545203942, "darwin": 0.008302602178538798}, "confidence": 0.9037339241483229}}, "usage": {"input_tokens": 653, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:19Z"}
{"id": "c-socrates-01-dessert-02", "uid": "c095", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "prompt_sha256_by_order": ["1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191"], "choice": "socrates", "probs": {"darwin": 0.009640945752902084, "einstein": 0.014347902331394945, "hypatia": 0.048130382958656265, "ibn_sina": 0.01103789090519898, "sagan": 0.01077231855471254, "socrates": 0.9060705594971352}, "confidence": 0.7885804966842395, "seconds": 0.18657156100046754, "seconds_all_orders": 1.1194293660028052, "seconds_by_order": [0.18605722800020885, 0.1866460340006597, 0.1864153360002092, 0.18699541100068018, 0.1871836890004488, 0.18613166800059844], "argmax_stable": true, "spread": {"darwin": 0.02088266449610898, "einstein": 0.007368742394323155, "hypatia": 0.0220694008193328, "ibn_sina": 0.011773028011132544, "sagan": 0.0038232038817475203, "socrates": 0.0382959200505395}, "spread_2dp": {"darwin": 0.019999999999999997, "einstein": 0.0, "hypatia": 0.020000000000000004, "ibn_sina": 0.01, "sagan": 0.0, "socrates": 0.040000000000000036}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.009640945752902084, "einstein": 0.014347902331394945, "hypatia": 0.048130382958656265, "ibn_sina": 0.01103789090519898, "sagan": 0.01077231855471254, "socrates": 0.9060705594971352}, {"ibn_sina": 0.007732530105113733, "hypatia": 0.026060982139323466, "sagan": 0.007924398940292815, "socrates": 0.9222187939987294, "darwin": 0.02908413487946874, "einstein": 0.00697915993707179}, {"hypatia": 0.04127473879011372, "darwin": 0.017356138363521345, "sagan": 0.006949114672965019, "ibn_sina": 0.009720044317732561, "einstein": 0.009330239921098955, "socrates": 0.9153697239345684}, {"ibn_sina": 0.006925741520912354, "einstein": 0.010349443272528236, "darwin": 0.012139267879164327, "sagan": 0.00899489454726561, "hypatia": 0.033261338642637195, "socrates": 0.9283293141374922}, {"einstein": 0.010664233127610659, "sagan": 0.00973044048272478, "socrates": 0.8900333940869527, "ibn_sina": 0.018698769532044898, "darwin": 0.026731546909907172, "hypatia": 0.044141615860759834}, {"socrates": 0.9155802908548705, "ibn_sina": 0.00994610218203553, "hypatia": 0.047196920969944635, "einstein": 0.009116280408025892, "sagan": 0.009958935201763684, "darwin": 0.00820147038335976}], "prompt_tokens": 511, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.009640945752902084, "einstein": 0.014347902331394945, "hypatia": 0.048130382958656265, "ibn_sina": 0.01103789090519898, "sagan": 0.01077231855471254, "socrates": 0.9060705594971352}, "confidence": 0.7885804966842395}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.007732530105113733, "hypatia": 0.026060982139323466, "sagan": 0.007924398940292815, "socrates": 0.9222187939987294, "darwin": 0.02908413487946874, "einstein": 0.00697915993707179}, "confidence": 0.8226206349880514}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.04127473879011372, "darwin": 0.017356138363521345, "sagan": 0.006949114672965019, "ibn_sina": 0.009720044317732561, "einstein": 0.009330239921098955, "socrates": 0.9153697239345684}, "confidence": 0.8081636727164749}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.006925741520912354, "einstein": 0.010349443272528236, "darwin": 0.012139267879164327, "sagan": 0.00899489454726561, "hypatia": 0.033261338642637195, "socrates": 0.9283293141374922}, "confidence": 0.8359419747514135}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.010664233127610659, "sagan": 0.00973044048272478, "socrates": 0.8900333940869527, "ibn_sina": 0.018698769532044898, "darwin": 0.026731546909907172, "hypatia": 0.044141615860759834}, "confidence": 0.754456662115764}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.9155802908548705, "ibn_sina": 0.00994610218203553, "hypatia": 0.047196920969944635, "einstein": 0.009116280408025892, "sagan": 0.009958935201763684, "darwin": 0.00820147038335976}, "confidence": 0.8090359532497933}}, "usage": {"input_tokens": 511, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:21Z"}
{"id": "c-socrates-01-main-01", "uid": "c096", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "prompt_sha256_by_order": ["d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79"], "choice": "socrates", "probs": {"darwin": 0.04204454496182124, "einstein": 0.04431301611275382, "hypatia": 0.005466261550139082, "ibn_sina": 0.0098207228872607, "sagan": 0.01862593224026494, "socrates": 0.8797295222477602}, "confidence": 0.7337544057250123, "seconds": 0.19220432599953105, "seconds_all_orders": 1.1532259559971862, "seconds_by_order": [0.19214101799843775, 0.19182508099947881, 0.1923924860002444, 0.19277589299963438, 0.19195430899890198, 0.19213716900048894], "argmax_stable": true, "spread": {"darwin": 0.04656010899930145, "einstein": 0.023168160563755435, "hypatia": 0.0021555975394672664, "ibn_sina": 0.004417835513744663, "sagan": 0.009523625388379288, "socrates": 0.03907726263473377}, "spread_2dp": {"darwin": 0.049999999999999996, "einstein": 0.02, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.01, "socrates": 0.040000000000000036}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.04204454496182124, "einstein": 0.04431301611275382, "hypatia": 0.005466261550139082, "ibn_sina": 0.0098207228872607, "sagan": 0.01862593224026494, "socrates": 0.8797295222477602}, {"ibn_sina": 0.01286263790808238, "hypatia": 0.005625371967619368, "sagan": 0.016735148825783603, "socrates": 0.8643893012997949, "darwin": 0.06794393884632097, "einstein": 0.03244360115239882}, {"hypatia": 0.006249386569854497, "darwin": 0.08860465396112269, "sagan": 0.013346722395229625, "ibn_sina": 0.008537983918362746, "einstein": 0.027740379633867714, "socrates": 0.8555208735215628}, {"ibn_sina": 0.010555408894435044, "einstein": 0.02425302930457092, "darwin": 0.053302127149213376, "sagan": 0.012457410778720035, "hypatia": 0.006785675640693582, "socrates": 0.892646348232367}, {"einstein": 0.036296596569388564, "sagan": 0.01991897084246131, "socrates": 0.8535690855976332, "ibn_sina": 0.011257411755775798, "darwin": 0.0713360761451347, "hypatia": 0.007621859089606348}, {"socrates": 0.8808826085775479, "ibn_sina": 0.01295581943210741, "hypatia": 0.006122773972180098, "einstein": 0.021144855548998382, "sagan": 0.021981036167099323, "darwin": 0.05691290630206697}], "prompt_tokens": 553, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.04204454496182124, "einstein": 0.04431301611275382, "hypatia": 0.005466261550139082, "ibn_sina": 0.0098207228872607, "sagan": 0.01862593224026494, "socrates": 0.8797295222477602}, "confidence": 0.7337544057250123}}, "usage": {"input_tokens": 553, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.01286263790808238, "hypatia": 0.005625371967619368, "sagan": 0.016735148825783603, "socrates": 0.8643893012997949, "darwin": 0.06794393884632097, "einstein": 0.03244360115239882}, "confidence": 0.7039779853038228}}, "usage": {"input_tokens": 553, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.006249386569854497, "darwin": 0.08860465396112269, "sagan": 0.013346722395229625, "ibn_sina": 0.008537983918362746, "einstein": 0.027740379633867714, "socrates": 0.8555208735215628}, "confidence": 0.6889916384769812}}, "usage": {"input_tokens": 553, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.010555408894435044, "einstein": 0.02425302930457092, "darwin": 0.053302127149213376, "sagan": 0.012457410778720035, "hypatia": 0.006785675640693582, "socrates": 0.892646348232367}, "confidence": 0.7606713740029585}}, "usage": {"input_tokens": 553, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.036296596569388564, "sagan": 0.01991897084246131, "socrates": 0.8535690855976332, "ibn_sina": 0.011257411755775798, "darwin": 0.0713360761451347, "hypatia": 0.007621859089606348}, "confidence": 0.6826816600301546}}, "usage": {"input_tokens": 553, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.8808826085775479, "ibn_sina": 0.01295581943210741, "hypatia": 0.006122773972180098, "einstein": 0.021144855548998382, "sagan": 0.021981036167099323, "darwin": 0.05691290630206697}, "confidence": 0.7363946337802869}}, "usage": {"input_tokens": 553, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:22Z"}
{"id": "c-socrates-01-main-02", "uid": "c097", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "prompt_sha256_by_order": ["b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d"], "choice": "socrates", "probs": {"darwin": 0.02897912014373259, "einstein": 0.01667227391316982, "hypatia": 0.01180201339088583, "ibn_sina": 0.04505683685832107, "sagan": 0.006368033589748996, "socrates": 0.8911217221041416}, "confidence": 0.7569107627765774, "seconds": 0.2540339390000857, "seconds_all_orders": 1.5242036340005143, "seconds_by_order": [0.25332540400086145, 0.25413554800070415, 0.254776422998475, 0.2539900399988255, 0.25339105500097503, 0.2545851640006731], "argmax_stable": true, "spread": {"darwin": 0.01185166947095236, "einstein": 0.009443992686659236, "hypatia": 0.013076628879983114, "ibn_sina": 0.03806275523803557, "sagan": 0.002655567616068136, "socrates": 0.04160000775937567}, "spread_2dp": {"darwin": 0.009999999999999998, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.04000000000000001, "sagan": 0.01, "socrates": 0.050000000000000044}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.02897912014373259, "einstein": 0.01667227391316982, "hypatia": 0.01180201339088583, "ibn_sina": 0.04505683685832107, "sagan": 0.006368033589748996, "socrates": 0.8911217221041416}, {"ibn_sina": 0.030633516359835078, "hypatia": 0.014169128205342626, "sagan": 0.006671489417646431, "socrates": 0.9057863514325508, "darwin": 0.03173418944395967, "einstein": 0.01100532514066533}, {"hypatia": 0.022879612096456975, "darwin": 0.020201223875570048, "sagan": 0.005265605124530935, "ibn_sina": 0.04863490870845107, "einstein": 0.0163853370316067, "socrates": 0.8866333131633842}, {"ibn_sina": 0.028656875433623274, "einstein": 0.008297444531474008, "darwin": 0.019993381797568134, "sagan": 0.0059090141476272435, "hypatia": 0.011762743905193742, "socrates": 0.9253805401845137}, {"einstein": 0.011581467026465005, "sagan": 0.005549866744465056, "socrates": 0.9088281915653544, "ibn_sina": 0.031111837276769132, "darwin": 0.01988251997300731, "hypatia": 0.023046117413939145}, {"socrates": 0.883780532425138, "ibn_sina": 0.06671963067165884, "hypatia": 0.00996948853395603, "einstein": 0.0072282812265105825, "sagan": 0.0040159218015782955, "darwin": 0.02828614534115819}], "prompt_tokens": 695, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.02897912014373259, "einstein": 0.01667227391316982, "hypatia": 0.01180201339088583, "ibn_sina": 0.04505683685832107, "sagan": 0.006368033589748996, "socrates": 0.8911217221041416}, "confidence": 0.7569107627765774}}, "usage": {"input_tokens": 695, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.030633516359835078, "hypatia": 0.014169128205342626, "sagan": 0.006671489417646431, "socrates": 0.9057863514325508, "darwin": 0.03173418944395967, "einstein": 0.01100532514066533}, "confidence": 0.7873129308307922}}, "usage": {"input_tokens": 695, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.022879612096456975, "darwin": 0.020201223875570048, "sagan": 0.005265605124530935, "ibn_sina": 0.04863490870845107, "einstein": 0.0163853370316067, "socrates": 0.8866333131633842}, "confidence": 0.7476541099826575}}, "usage": {"input_tokens": 695, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.028656875433623274, "einstein": 0.008297444531474008, "darwin": 0.019993381797568134, "sagan": 0.0059090141476272435, "hypatia": 0.011762743905193742, "socrates": 0.9253805401845137}, "confidence": 0.8293506665867593}}, "usage": {"input_tokens": 695, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.011581467026465005, "sagan": 0.005549866744465056, "socrates": 0.9088281915653544, "ibn_sina": 0.031111837276769132, "darwin": 0.01988251997300731, "hypatia": 0.023046117413939145}, "confidence": 0.793633597276467}}, "usage": {"input_tokens": 695, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.883780532425138, "ibn_sina": 0.06671963067165884, "hypatia": 0.00996948853395603, "einstein": 0.0072282812265105825, "sagan": 0.0040159218015782955, "darwin": 0.02828614534115819}, "confidence": 0.7437848932095024}}, "usage": {"input_tokens": 695, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:23Z"}
{"id": "c-socrates-01-main-02", "uid": "c098", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "prompt_sha256_by_order": ["7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53"], "choice": "ibn_sina", "probs": {"darwin": 0.09134791182465796, "einstein": 0.09524924656708951, "hypatia": 0.04789503648476202, "ibn_sina": 0.4297612090459203, "sagan": 0.09060482371887837, "socrates": 0.24514177235869183}, "confidence": 0.12725105670772158, "seconds": 0.18579146100000798, "seconds_all_orders": 1.114748766000048, "seconds_by_order": [0.18612209799903212, 0.18554499200035934, 0.18520089599951461, 0.18560745300055714, 0.18596426000112842, 0.18630906699945626], "argmax_stable": false, "spread": {"darwin": 0.02992847274432714, "einstein": 0.018844500198622716, "hypatia": 0.03173942629226798, "ibn_sina": 0.11716549807370952, "sagan": 0.06648712358632575, "socrates": 0.17700193221637683}, "spread_2dp": {"darwin": 0.03, "einstein": 0.020000000000000004, "hypatia": 0.03, "ibn_sina": 0.12, "sagan": 0.06, "socrates": 0.16999999999999998}, "choices_by_order": ["ibn_sina", "socrates", "socrates", "socrates", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.09134791182465796, "einstein": 0.09524924656708951, "hypatia": 0.04789503648476202, "ibn_sina": 0.4297612090459203, "sagan": 0.09060482371887837, "socrates": 0.24514177235869183}, {"ibn_sina": 0.31636075814380016, "hypatia": 0.05131473666668908, "sagan": 0.06531123610277001, "socrates": 0.3813812914345785, "darwin": 0.09027155281134935, "einstein": 0.09536042484081286}, {"hypatia": 0.06800560089607326, "darwin": 0.06704476742772353, "sagan": 0.058482657909511036, "ibn_sina": 0.3125957109722108, "einstein": 0.08139812519415277, "socrates": 0.41247313760032867}, {"ibn_sina": 0.32234528698356, "einstein": 0.07651592464219015, "darwin": 0.06461149400080182, "sagan": 0.06130027398361652, "hypatia": 0.05308331581476279, "socrates": 0.42214370457506867}, {"einstein": 0.0836823497885599, "sagan": 0.12496978149583679, "socrates": 0.28498934462517833, "ibn_sina": 0.3653046222330642, "darwin": 0.06141943908033082, "hypatia": 0.07963446277703}, {"socrates": 0.2845850508068823, "ibn_sina": 0.41790946413049174, "hypatia": 0.06086123229528065, "einstein": 0.08402714446548336, "sagan": 0.07705194800665369, "darwin": 0.07556516029520818}], "prompt_tokens": 501, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.09134791182465796, "einstein": 0.09524924656708951, "hypatia": 0.04789503648476202, "ibn_sina": 0.4297612090459203, "sagan": 0.09060482371887837, "socrates": 0.24514177235869183}, "confidence": 0.12725105670772158}}, "usage": {"input_tokens": 501, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.31636075814380016, "hypatia": 0.05131473666668908, "sagan": 0.06531123610277001, "socrates": 0.3813812914345785, "darwin": 0.09027155281134935, "einstein": 0.09536042484081286}, "confidence": 0.12361257085943547}}, "usage": {"input_tokens": 501, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.06800560089607326, "darwin": 0.06704476742772353, "sagan": 0.058482657909511036, "ibn_sina": 0.3125957109722108, "einstein": 0.08139812519415277, "socrates": 0.41247313760032867}, "confidence": 0.1444189676968468}}, "usage": {"input_tokens": 501, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.32234528698356, "einstein": 0.07651592464219015, "darwin": 0.06461149400080182, "sagan": 0.06130027398361652, "hypatia": 0.05308331581476279, "socrates": 0.42214370457506867}, "confidence": 0.1584600222905174}}, "usage": {"input_tokens": 501, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.0836823497885599, "sagan": 0.12496978149583679, "socrates": 0.28498934462517833, "ibn_sina": 0.3653046222330642, "darwin": 0.06141943908033082, "hypatia": 0.07963446277703}, "confidence": 0.09688068482405336}}, "usage": {"input_tokens": 501, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.2845850508068823, "ibn_sina": 0.41790946413049174, "hypatia": 0.06086123229528065, "einstein": 0.08402714446548336, "sagan": 0.07705194800665369, "darwin": 0.07556516029520818}, "confidence": 0.13365846171780932}}, "usage": {"input_tokens": 501, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:24Z"}
{"id": "c-socrates-01-main-02", "uid": "c099", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "prompt_sha256_by_order": ["d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755"], "choice": "socrates", "probs": {"darwin": 0.005821688567236369, "einstein": 0.006472467940036578, "hypatia": 0.030641964433358888, "ibn_sina": 0.013223303112563364, "sagan": 0.008286714775644122, "socrates": 0.9355538611711607}, "confidence": 0.8518231209071091, "seconds": 0.22667928466671583, "seconds_all_orders": 1.360075708000295, "seconds_by_order": [0.2265574650009512, 0.22642674700000498, 0.2265411859989399, 0.22728301999995892, 0.227119371000299, 0.22614791900014097], "argmax_stable": true, "spread": {"darwin": 0.003803526572758533, "einstein": 0.0026163395788546253, "hypatia": 0.014053474856762428, "ibn_sina": 0.012965148054766878, "sagan": 0.0029326273662464604, "socrates": 0.016051526314876163}, "spread_2dp": {"darwin": 0.0, "einstein": 0.01, "hypatia": 0.010000000000000002, "ibn_sina": 0.01, "sagan": 0.0, "socrates": 0.009999999999999898}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.005821688567236369, "einstein": 0.006472467940036578, "hypatia": 0.030641964433358888, "ibn_sina": 0.013223303112563364, "sagan": 0.008286714775644122, "socrates": 0.9355538611711607}, {"ibn_sina": 0.012683487975157233, "hypatia": 0.03098518827063665, "sagan": 0.008014529198830958, "socrates": 0.9324509636515937, "darwin": 0.009111391397095785, "einstein": 0.006754439506685667}, {"hypatia": 0.030258843791270147, "darwin": 0.005307864824337252, "sagan": 0.005354087409397662, "ibn_sina": 0.019485096078955814, "einstein": 0.006865467181577201, "socrates": 0.9327286407144619}, {"ibn_sina": 0.0065199480241889375, "einstein": 0.004668006860423622, "darwin": 0.0055748897447479755, "sagan": 0.007133679067522733, "hypatia": 0.03429994651730684, "socrates": 0.94180352978581}, {"einstein": 0.007284346439278247, "sagan": 0.008063107010951018, "socrates": 0.927284966301237, "ibn_sina": 0.012227974013451666, "darwin": 0.0069798079353547565, "hypatia": 0.03815979829972717}, {"socrates": 0.9257520034709338, "ibn_sina": 0.012109742174095089, "hypatia": 0.044312318648032575, "einstein": 0.005955502666984134, "sagan": 0.006316675356278419, "darwin": 0.005553757683676018}], "prompt_tokens": 641, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.005821688567236369, "einstein": 0.006472467940036578, "hypatia": 0.030641964433358888, "ibn_sina": 0.013223303112563364, "sagan": 0.008286714775644122, "socrates": 0.9355538611711607}, "confidence": 0.8518231209071091}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.012683487975157233, "hypatia": 0.03098518827063665, "sagan": 0.008014529198830958, "socrates": 0.9324509636515937, "darwin": 0.009111391397095785, "einstein": 0.006754439506685667}, "confidence": 0.8449343499504257}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.030258843791270147, "darwin": 0.005307864824337252, "sagan": 0.005354087409397662, "ibn_sina": 0.019485096078955814, "einstein": 0.006865467181577201, "socrates": 0.9327286407144619}, "confidence": 0.8456583497517288}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.0065199480241889375, "einstein": 0.004668006860423622, "darwin": 0.0055748897447479755, "sagan": 0.007133679067522733, "hypatia": 0.03429994651730684, "socrates": 0.94180352978581}, "confidence": 0.8659799725973112}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.007284346439278247, "sagan": 0.008063107010951018, "socrates": 0.927284966301237, "ibn_sina": 0.012227974013451666, "darwin": 0.0069798079353547565, "hypatia": 0.03815979829972717}, "confidence": 0.8339558744794813}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.9257520034709338, "ibn_sina": 0.012109742174095089, "hypatia": 0.044312318648032575, "einstein": 0.005955502666984134, "sagan": 0.006316675356278419, "darwin": 0.005553757683676018}, "confidence": 0.8310798543927012}}, "usage": {"input_tokens": 641, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:26Z"}
{"id": "c-socrates-01-main-02", "uid": "c100", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "prompt_sha256_by_order": ["b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2"], "choice": "ibn_sina", "probs": {"darwin": 0.06977349367266597, "einstein": 0.09599813032374578, "hypatia": 0.043874536599996, "ibn_sina": 0.5704288473725019, "sagan": 0.11576653609480476, "socrates": 0.10415845593628564}, "confidence": 0.23877868137274091, "seconds": 0.18563467700035594, "seconds_all_orders": 1.1138080620021356, "seconds_by_order": [0.18513193600119848, 0.1864855149997311, 0.186088837999705, 0.18611236800097686, 0.18515390600077808, 0.18483549899974605], "argmax_stable": true, "spread": {"darwin": 0.013970068553554565, "einstein": 0.04404283055937079, "hypatia": 0.03503434651883526, "ibn_sina": 0.273925715882817, "sagan": 0.05471456233517071, "socrates": 0.23490321514951035}, "spread_2dp": {"darwin": 0.010000000000000009, "einstein": 0.05, "hypatia": 0.030000000000000006, "ibn_sina": 0.27, "sagan": 0.06, "socrates": 0.24000000000000002}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.06977349367266597, "einstein": 0.09599813032374578, "hypatia": 0.043874536599996, "ibn_sina": 0.5704288473725019, "sagan": 0.11576653609480476, "socrates": 0.10415845593628564}, {"ibn_sina": 0.41045626089051696, "hypatia": 0.05799047273465095, "sagan": 0.08510740620410774, "socrates": 0.3236648416670593, "darwin": 0.05938745074542965, "einstein": 0.06339356775823544}, {"hypatia": 0.06239754081953467, "darwin": 0.06640572549391088, "sagan": 0.06105197375963405, "ibn_sina": 0.6057288477660462, "einstein": 0.07211034331489974, "socrates": 0.13230556884597444}, {"ibn_sina": 0.367862904837557, "einstein": 0.08972157297328859, "darwin": 0.06901774246215103, "sagan": 0.06327605532171317, "hypatia": 0.07106005331949419, "socrates": 0.339061671085796}, {"einstein": 0.07236922836528142, "sagan": 0.10733456241003556, "socrates": 0.15226494460389078, "ibn_sina": 0.5511278614857246, "darwin": 0.055803425119111405, "hypatia": 0.06109997801595625}, {"socrates": 0.11985256428136404, "ibn_sina": 0.641788620720374, "hypatia": 0.03602570680065893, "einstein": 0.05195529976437499, "sagan": 0.08777155899468833, "darwin": 0.06260624943853974}], "prompt_tokens": 503, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"darwin": 0.06977349367266597, "einstein": 0.09599813032374578, "hypatia": 0.043874536599996, "ibn_sina": 0.5704288473725019, "sagan": 0.11576653609480476, "socrates": 0.10415845593628564}, "confidence": 0.23877868137274091}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.41045626089051696, "hypatia": 0.05799047273465095, "sagan": 0.08510740620410774, "socrates": 0.3236648416670593, "darwin": 0.05938745074542965, "einstein": 0.06339356775823544}, "confidence": 0.14966206131222978}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"hypatia": 0.06239754081953467, "darwin": 0.06640572549391088, "sagan": 0.06105197375963405, "ibn_sina": 0.6057288477660462, "einstein": 0.07211034331489974, "socrates": 0.13230556884597444}, "confidence": 0.2819711429863321}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"ibn_sina": 0.367862904837557, "einstein": 0.08972157297328859, "darwin": 0.06901774246215103, "sagan": 0.06327605532171317, "hypatia": 0.07106005331949419, "socrates": 0.339061671085796}, "confidence": 0.12658328000948424}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"einstein": 0.07236922836528142, "sagan": 0.10733456241003556, "socrates": 0.15226494460389078, "ibn_sina": 0.5511278614857246, "darwin": 0.055803425119111405, "hypatia": 0.06109997801595625}, "confidence": 0.22063813135788238}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "ibn_sina", "probabilities": {"socrates": 0.11985256428136404, "ibn_sina": 0.641788620720374, "hypatia": 0.03602570680065893, "einstein": 0.05195529976437499, "sagan": 0.08777155899468833, "darwin": 0.06260624943853974}, "confidence": 0.33025343753462855}}, "usage": {"input_tokens": 503, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:27Z"}
{"id": "c-socrates-01-main-03", "uid": "c101", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "prompt_sha256_by_order": ["087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b"], "choice": "socrates", "probs": {"darwin": 0.010260301520915959, "einstein": 0.010381968362129509, "hypatia": 0.24141288593919852, "ibn_sina": 0.02827622431284359, "sagan": 0.1418406484425869, "socrates": 0.5678279714223256}, "confidence": 0.2822081921121543, "seconds": 0.22187280516664032, "seconds_all_orders": 1.331236830999842, "seconds_by_order": [0.2222307699994417, 0.22231445000034, 0.22147009699983755, 0.22152069699950516, 0.2216662950013415, 0.222034521999376], "argmax_stable": false, "spread": {"darwin": 0.003981527554726087, "einstein": 0.0019636321604978475, "hypatia": 0.23281822071628977, "ibn_sina": 0.011509620075216772, "sagan": 0.1791963109296635, "socrates": 0.2663489424367962}, "spread_2dp": {"darwin": 0.0, "einstein": 0.0, "hypatia": 0.22999999999999998, "ibn_sina": 0.009999999999999998, "sagan": 0.18, "socrates": 0.25999999999999995}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "hypatia", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.010260301520915959, "einstein": 0.010381968362129509, "hypatia": 0.24141288593919852, "ibn_sina": 0.02827622431284359, "sagan": 0.1418406484425869, "socrates": 0.5678279714223256}, {"ibn_sina": 0.01959127429027095, "hypatia": 0.2996180060877478, "sagan": 0.15172192901339251, "socrates": 0.506572870436317, "darwin": 0.012773044345021786, "einstein": 0.009722875827249958}, {"hypatia": 0.3363206310761852, "darwin": 0.011721703880257867, "sagan": 0.1452716951259281, "ibn_sina": 0.017747413860965434, "einstein": 0.011686507987747806, "socrates": 0.4772520480689155}, {"ibn_sina": 0.02291370934212092, "einstein": 0.010444809899429992, "darwin": 0.012901094028050314, "sagan": 0.2122268090036228, "hypatia": 0.16966557850090727, "socrates": 0.5718479992258687}, {"einstein": 0.010225663734656413, "sagan": 0.28042073136962964, "socrates": 0.3054990567890725, "ibn_sina": 0.0202799425229655, "darwin": 0.009041829607228519, "hypatia": 0.37453277597644735}, {"socrates": 0.4963614512374288, "ibn_sina": 0.01676660423762682, "hypatia": 0.14171455526015758, "einstein": 0.011097072730581747, "sagan": 0.3210369593722504, "darwin": 0.013023357161954605}], "prompt_tokens": 587, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.010260301520915959, "einstein": 0.010381968362129509, "hypatia": 0.24141288593919852, "ibn_sina": 0.02827622431284359, "sagan": 0.1418406484425869, "socrates": 0.5678279714223256}, "confidence": 0.2822081921121543}}, "usage": {"input_tokens": 587, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.01959127429027095, "hypatia": 0.2996180060877478, "sagan": 0.15172192901339251, "socrates": 0.506572870436317, "darwin": 0.012773044345021786, "einstein": 0.009722875827249958}, "confidence": 0.2440576832585691}}, "usage": {"input_tokens": 587, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.3363206310761852, "darwin": 0.011721703880257867, "sagan": 0.1452716951259281, "ibn_sina": 0.017747413860965434, "einstein": 0.011686507987747806, "socrates": 0.4772520480689155}, "confidence": 0.23508867182533005}}, "usage": {"input_tokens": 587, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.02291370934212092, "einstein": 0.010444809899429992, "darwin": 0.012901094028050314, "sagan": 0.2122268090036228, "hypatia": 0.16966557850090727, "socrates": 0.5718479992258687}, "confidence": 0.28196479787596945}}, "usage": {"input_tokens": 587, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "hypatia", "probabilities": {"einstein": 0.010225663734656413, "sagan": 0.28042073136962964, "socrates": 0.3054990567890725, "ibn_sina": 0.0202799425229655, "darwin": 0.009041829607228519, "hypatia": 0.37453277597644735}, "confidence": 0.1754054266140518}}, "usage": {"input_tokens": 587, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.4963614512374288, "ibn_sina": 0.01676660423762682, "hypatia": 0.14171455526015758, "einstein": 0.011097072730581747, "sagan": 0.3210369593722504, "darwin": 0.013023357161954605}, "confidence": 0.244115567923249}}, "usage": {"input_tokens": 587, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:28Z"}
{"id": "c-socrates-01-main-03", "uid": "c102", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "prompt_sha256_by_order": ["da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb"], "choice": "socrates", "probs": {"darwin": 0.018090428701741337, "einstein": 0.014799556388459556, "hypatia": 0.015865874446927886, "ibn_sina": 0.0981068932309685, "sagan": 0.013467640523971779, "socrates": 0.839669606707931}, "confidence": 0.6587792856659689, "seconds": 0.22895368216662368, "seconds_all_orders": 1.373722092999742, "seconds_by_order": [0.22928275300000678, 0.22863562800012005, 0.228205691999392, 0.22921273300016765, 0.22925575299996126, 0.22912953400009428], "argmax_stable": true, "spread": {"darwin": 0.009900160193053092, "einstein": 0.019996276263270878, "hypatia": 0.03367690955355511, "ibn_sina": 0.1205523375469359, "sagan": 0.00661557880402566, "socrates": 0.16957858843606322}, "spread_2dp": {"darwin": 0.01, "einstein": 0.019999999999999997, "hypatia": 0.04, "ibn_sina": 0.13, "sagan": 0.0, "socrates": 0.17000000000000004}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.018090428701741337, "einstein": 0.014799556388459556, "hypatia": 0.015865874446927886, "ibn_sina": 0.0981068932309685, "sagan": 0.013467640523971779, "socrates": 0.839669606707931}, {"ibn_sina": 0.0837326185469313, "hypatia": 0.019715721632149904, "sagan": 0.010323467131621047, "socrates": 0.8423465501909158, "darwin": 0.01665976990747123, "einstein": 0.02722187259091066}, {"hypatia": 0.027643039911118196, "darwin": 0.012440261159006073, "sagan": 0.009088846096355368, "ibn_sina": 0.0647699586208611, "einstein": 0.02006364128998129, "socrates": 0.8659942529226778}, {"ibn_sina": 0.06540104477088268, "einstein": 0.012040644766765468, "darwin": 0.008190268508688245, "sagan": 0.008586057330016965, "hypatia": 0.015293359842743024, "socrates": 0.8904886247809036}, {"einstein": 0.023260074583983836, "sagan": 0.010029254203155358, "socrates": 0.7209100363448404, "ibn_sina": 0.185322296167797, "darwin": 0.01480310105296415, "hypatia": 0.045675237647259226}, {"socrates": 0.8460345724689284, "ibn_sina": 0.11711328146029609, "hypatia": 0.011998328093704115, "einstein": 0.007225596327639781, "sagan": 0.006852061719946119, "darwin": 0.010776159929485388}], "prompt_tokens": 663, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.018090428701741337, "einstein": 0.014799556388459556, "hypatia": 0.015865874446927886, "ibn_sina": 0.0981068932309685, "sagan": 0.013467640523971779, "socrates": 0.839669606707931}, "confidence": 0.6587792856659689}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.0837326185469313, "hypatia": 0.019715721632149904, "sagan": 0.010323467131621047, "socrates": 0.8423465501909158, "darwin": 0.01665976990747123, "einstein": 0.02722187259091066}, "confidence": 0.6616872687532991}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.027643039911118196, "darwin": 0.012440261159006073, "sagan": 0.009088846096355368, "ibn_sina": 0.0647699586208611, "einstein": 0.02006364128998129, "socrates": 0.8659942529226778}, "confidence": 0.7066542978559159}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.06540104477088268, "einstein": 0.012040644766765468, "darwin": 0.008190268508688245, "sagan": 0.008586057330016965, "hypatia": 0.015293359842743024, "socrates": 0.8904886247809036}, "confidence": 0.7573203428580471}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.023260074583983836, "sagan": 0.010029254203155358, "socrates": 0.7209100363448404, "ibn_sina": 0.185322296167797, "darwin": 0.01480310105296415, "hypatia": 0.045675237647259226}, "confidence": 0.46840313212487866}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.8460345724689284, "ibn_sina": 0.11711328146029609, "hypatia": 0.011998328093704115, "einstein": 0.007225596327639781, "sagan": 0.006852061719946119, "darwin": 0.010776159929485388}, "confidence": 0.675819116798866}}, "usage": {"input_tokens": 663, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:30Z"}
{"id": "c-socrates-01-starter-01", "uid": "c103", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "prompt_sha256_by_order": ["ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c"], "choice": "socrates", "probs": {"darwin": 0.006962384986920676, "einstein": 0.005115538532641881, "hypatia": 0.03048186992103159, "ibn_sina": 0.03611031664582413, "sagan": 0.008485278706277112, "socrates": 0.9128446112073046}, "confidence": 0.8027980324795212, "seconds": 0.18784688700028104, "seconds_all_orders": 1.1270813220016862, "seconds_by_order": [0.18782507400101167, 0.18696133100092993, 0.18795906299965282, 0.18827048999992257, 0.1880689320005331, 0.18799643199963612], "argmax_stable": true, "spread": {"darwin": 0.002396131271069201, "einstein": 0.00292841115637506, "hypatia": 0.018657704316021317, "ibn_sina": 0.02656383728513599, "sagan": 0.004968586953997831, "socrates": 0.0334411684948146}, "spread_2dp": {"darwin": 0.0, "einstein": 0.01, "hypatia": 0.020000000000000004, "ibn_sina": 0.020000000000000004, "sagan": 0.0, "socrates": 0.030000000000000027}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.006962384986920676, "einstein": 0.005115538532641881, "hypatia": 0.03048186992103159, "ibn_sina": 0.03611031664582413, "sagan": 0.008485278706277112, "socrates": 0.9128446112073046}, {"ibn_sina": 0.036017548336972766, "hypatia": 0.03179340332906079, "sagan": 0.0104829778427049, "socrates": 0.9079151106003309, "darwin": 0.0072428404690405524, "einstein": 0.0065481194218901155}, {"hypatia": 0.03199519112190674, "darwin": 0.007589213555726405, "sagan": 0.008699726619221127, "ibn_sina": 0.04933848802643193, "einstein": 0.007614898606420507, "socrates": 0.8947624820702933}, {"ibn_sina": 0.027249284373563484, "einstein": 0.004686487450045447, "darwin": 0.006325609431732, "sagan": 0.007026933929993386, "hypatia": 0.045804866070706156, "socrates": 0.9089068187439595}, {"einstein": 0.007501375261702895, "sagan": 0.0074354058924591335, "socrates": 0.8892260609870557, "ibn_sina": 0.053813121658699475, "darwin": 0.00843160676538489, "hypatia": 0.03359242943469798}, {"socrates": 0.9226672294818703, "ibn_sina": 0.032936538952853534, "hypatia": 0.02714716175468484, "einstein": 0.005699203427568623, "sagan": 0.005514390888707069, "darwin": 0.006035475494315689}], "prompt_tokens": 441, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.006962384986920676, "einstein": 0.005115538532641881, "hypatia": 0.03048186992103159, "ibn_sina": 0.03611031664582413, "sagan": 0.008485278706277112, "socrates": 0.9128446112073046}, "confidence": 0.8027980324795212}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.036017548336972766, "hypatia": 0.03179340332906079, "sagan": 0.0104829778427049, "socrates": 0.9079151106003309, "darwin": 0.0072428404690405524, "einstein": 0.0065481194218901155}, "confidence": 0.7921877941244027}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.03199519112190674, "darwin": 0.007589213555726405, "sagan": 0.008699726619221127, "ibn_sina": 0.04933848802643193, "einstein": 0.007614898606420507, "socrates": 0.8947624820702933}, "confidence": 0.76509897527522}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.027249284373563484, "einstein": 0.004686487450045447, "darwin": 0.006325609431732, "sagan": 0.007026933929993386, "hypatia": 0.045804866070706156, "socrates": 0.9089068187439595}, "confidence": 0.7948762824563804}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.007501375261702895, "sagan": 0.0074354058924591335, "socrates": 0.8892260609870557, "ibn_sina": 0.053813121658699475, "darwin": 0.00843160676538489, "hypatia": 0.03359242943469798}, "confidence": 0.753915926560939}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.9226672294818703, "ibn_sina": 0.032936538952853534, "hypatia": 0.02714716175468484, "einstein": 0.005699203427568623, "sagan": 0.005514390888707069, "darwin": 0.006035475494315689}, "confidence": 0.8238831000883634}}, "usage": {"input_tokens": 441, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:31Z"}
{"id": "c-socrates-01-starter-02", "uid": "c104", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "prompt_sha256_by_order": ["caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d"], "choice": "socrates", "probs": {"darwin": 0.003555190847433464, "einstein": 0.0040743539088806255, "hypatia": 0.02920889303265719, "ibn_sina": 0.009089525649551292, "sagan": 0.005085959616573661, "socrates": 0.9489860769449039}, "confidence": 0.8818785518451568, "seconds": 0.18980834783360478, "seconds_all_orders": 1.1388500870016287, "seconds_by_order": [0.18907737399968028, 0.19004144500104303, 0.19006984599946009, 0.1899981660008052, 0.190037016000133, 0.1896262400005071], "argmax_stable": true, "spread": {"darwin": 0.003683274648917197, "einstein": 0.001757618120503264, "hypatia": 0.10996520636831567, "ibn_sina": 0.004918066179734785, "sagan": 0.002301996265977725, "socrates": 0.10150656912924227}, "spread_2dp": {"darwin": 0.01, "einstein": 0.0, "hypatia": 0.11000000000000001, "ibn_sina": 0.01, "sagan": 0.01, "socrates": 0.09999999999999998}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.003555190847433464, "einstein": 0.0040743539088806255, "hypatia": 0.02920889303265719, "ibn_sina": 0.009089525649551292, "sagan": 0.005085959616573661, "socrates": 0.9489860769449039}, {"ibn_sina": 0.007499507740153599, "hypatia": 0.049328724143779125, "sagan": 0.005195603282467497, "socrates": 0.9268783332718663, "darwin": 0.006845984887801542, "einstein": 0.004251846673931987}, {"hypatia": 0.04225226642001383, "darwin": 0.0038220455849642615, "sagan": 0.0028936070164897725, "ibn_sina": 0.007008470418832334, "einstein": 0.003301797179586599, "socrates": 0.9407218133801132}, {"ibn_sina": 0.004729547438686683, "einstein": 0.002494228553428723, "darwin": 0.0031627102388843446, "sagan": 0.0029599065523657144, "hypatia": 0.13917409940097286, "socrates": 0.8474795078156616}, {"einstein": 0.003992999586625766, "sagan": 0.004824636946977734, "socrates": 0.9061683596942871, "ibn_sina": 0.009647613618421468, "darwin": 0.004724073827083291, "hypatia": 0.0706423163266046}, {"socrates": 0.9365679264071869, "ibn_sina": 0.009398575484251133, "hypatia": 0.04171154492173551, "einstein": 0.00408413026894897, "sagan": 0.004168145011111301, "darwin": 0.004069677906766312}], "prompt_tokens": 537, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.003555190847433464, "einstein": 0.0040743539088806255, "hypatia": 0.02920889303265719, "ibn_sina": 0.009089525649551292, "sagan": 0.005085959616573661, "socrates": 0.9489860769449039}, "confidence": 0.8818785518451568}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.007499507740153599, "hypatia": 0.049328724143779125, "sagan": 0.005195603282467497, "socrates": 0.9268783332718663, "darwin": 0.006845984887801542, "einstein": 0.004251846673931987}, "confidence": 0.8340219404002263}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.04225226642001383, "darwin": 0.0038220455849642615, "sagan": 0.0028936070164897725, "ibn_sina": 0.007008470418832334, "einstein": 0.003301797179586599, "socrates": 0.9407218133801132}, "confidence": 0.8641909428436509}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.004729547438686683, "einstein": 0.002494228553428723, "darwin": 0.0031627102388843446, "sagan": 0.0029599065523657144, "hypatia": 0.13917409940097286, "socrates": 0.8474795078156616}, "confidence": 0.6851659596273441}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.003992999586625766, "sagan": 0.004824636946977734, "socrates": 0.9061683596942871, "ibn_sina": 0.009647613618421468, "darwin": 0.004724073827083291, "hypatia": 0.0706423163266046}, "confidence": 0.7915432569477352}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.9365679264071869, "ibn_sina": 0.009398575484251133, "hypatia": 0.04171154492173551, "einstein": 0.00408413026894897, "sagan": 0.004168145011111301, "darwin": 0.004069677906766312}, "confidence": 0.8548459393680415}}, "usage": {"input_tokens": 537, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:32Z"}
{"id": "c-socrates-01-starter-03", "uid": "c105", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "prompt_sha256_by_order": ["b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f"], "choice": "socrates", "probs": {"darwin": 0.03655312972657633, "einstein": 0.22466916604668485, "hypatia": 0.0059709302866236755, "ibn_sina": 0.03675851022161063, "sagan": 0.014110510956755206, "socrates": 0.6819377527617493}, "confidence": 0.42212489284999144, "seconds": 0.22093300666680685, "seconds_all_orders": 1.3255980400008411, "seconds_by_order": [0.22192984300090757, 0.2206868339999346, 0.22150371599855134, 0.21998596000048565, 0.2199679900004412, 0.22152369700052077], "argmax_stable": true, "spread": {"darwin": 0.11432724499198428, "einstein": 0.1519123098858493, "hypatia": 0.00430743997538469, "ibn_sina": 0.038477802162513286, "sagan": 0.0091726320473957, "socrates": 0.22768009643075526}, "spread_2dp": {"darwin": 0.12, "einstein": 0.15000000000000002, "hypatia": 0.0, "ibn_sina": 0.039999999999999994, "sagan": 0.01, "socrates": 0.2300000000000001}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.03655312972657633, "einstein": 0.22466916604668485, "hypatia": 0.0059709302866236755, "ibn_sina": 0.03675851022161063, "sagan": 0.014110510956755206, "socrates": 0.6819377527617493}, {"ibn_sina": 0.043291821941056366, "hypatia": 0.008070758168172968, "sagan": 0.010669751677117235, "socrates": 0.6889412067858313, "darwin": 0.13882128750394757, "einstein": 0.11020517392387459}, {"hypatia": 0.0075987419349398625, "darwin": 0.11911419058307297, "sagan": 0.006767300011351727, "ibn_sina": 0.02141935649988149, "einstein": 0.2621174838097239, "socrates": 0.5829829271610301}, {"ibn_sina": 0.023020951167457192, "einstein": 0.11342765081796788, "darwin": 0.03212954019854401, "sagan": 0.013486839191108617, "hypatia": 0.007271995033136833, "socrates": 0.8106630235917853}, {"einstein": 0.13676772103791746, "sagan": 0.015808055099714093, "socrates": 0.6309010099310145, "ibn_sina": 0.05989715866239478, "darwin": 0.1464567851905283, "hypatia": 0.010169270078430789}, {"socrates": 0.7657088072294691, "ibn_sina": 0.04162918663158, "hypatia": 0.005861830103046099, "einstein": 0.13582312178564138, "sagan": 0.015939932058747427, "darwin": 0.03503712219151599}], "prompt_tokens": 577, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.03655312972657633, "einstein": 0.22466916604668485, "hypatia": 0.0059709302866236755, "ibn_sina": 0.03675851022161063, "sagan": 0.014110510956755206, "socrates": 0.6819377527617493}, "confidence": 0.42212489284999144}}, "usage": {"input_tokens": 577, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.043291821941056366, "hypatia": 0.008070758168172968, "sagan": 0.010669751677117235, "socrates": 0.6889412067858313, "darwin": 0.13882128750394757, "einstein": 0.11020517392387459}, "confidence": 0.4097316150599191}}, "usage": {"input_tokens": 577, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.0075987419349398625, "darwin": 0.11911419058307297, "sagan": 0.006767300011351727, "ibn_sina": 0.02141935649988149, "einstein": 0.2621174838097239, "socrates": 0.5829829271610301}, "confidence": 0.30799022216746097}}, "usage": {"input_tokens": 577, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.023020951167457192, "einstein": 0.11342765081796788, "darwin": 0.03212954019854401, "sagan": 0.013486839191108617, "hypatia": 0.007271995033136833, "socrates": 0.8106630235917853}, "confidence": 0.6062049016938538}}, "usage": {"input_tokens": 577, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.13676772103791746, "sagan": 0.015808055099714093, "socrates": 0.6309010099310145, "ibn_sina": 0.05989715866239478, "darwin": 0.1464567851905283, "hypatia": 0.010169270078430789}, "confidence": 0.3305584744648192}}, "usage": {"input_tokens": 577, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.7657088072294691, "ibn_sina": 0.04162918663158, "hypatia": 0.005861830103046099, "einstein": 0.13582312178564138, "sagan": 0.015939932058747427, "darwin": 0.03503712219151599}, "confidence": 0.5296083153731657}}, "usage": {"input_tokens": 577, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:33Z"}
{"id": "c-socrates-01-starter-03", "uid": "c106", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "prompt_sha256_by_order": ["4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269"], "choice": "socrates", "probs": {"darwin": 0.012286566918535644, "einstein": 0.018454644591876835, "hypatia": 0.004471421570093671, "ibn_sina": 0.006220106501239739, "sagan": 0.010571965068651955, "socrates": 0.947995295349602}, "confidence": 0.8792284761037509, "seconds": 0.22426580766659754, "seconds_all_orders": 1.3455948459995852, "seconds_by_order": [0.22425622399896383, 0.2241502560009394, 0.22367030900022655, 0.22450736199971288, 0.22485747000064293, 0.22415322499909962], "argmax_stable": true, "spread": {"darwin": 0.008737935989341111, "einstein": 0.011715380846134823, "hypatia": 0.002779531714605101, "ibn_sina": 0.0036098462429695697, "sagan": 0.005469859902234385, "socrates": 0.013492723200638101}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.0, "socrates": 0.010000000000000009}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.012286566918535644, "einstein": 0.018454644591876835, "hypatia": 0.004471421570093671, "ibn_sina": 0.006220106501239739, "sagan": 0.010571965068651955, "socrates": 0.947995295349602}, {"ibn_sina": 0.007286653720488884, "hypatia": 0.005400448709889409, "sagan": 0.008391780119708158, "socrates": 0.9480445169247135, "darwin": 0.014867451932552466, "einstein": 0.016009148592647675}, {"hypatia": 0.0059405161975442, "darwin": 0.007050619482146086, "sagan": 0.00510210516641757, "ibn_sina": 0.007585503649862217, "einstein": 0.024245813678909663, "socrates": 0.9500754418251203}, {"ibn_sina": 0.0058129823172400775, "einstein": 0.01253043283277484, "darwin": 0.006281895951539681, "sagan": 0.0096534954257128, "hypatia": 0.007214918136148576, "socrates": 0.958506275336584}, {"einstein": 0.02072378515744136, "sagan": 0.008859452034017536, "socrates": 0.9450135521359458, "ibn_sina": 0.009422828560209647, "darwin": 0.008729428827686823, "hypatia": 0.007250953284698772}, {"socrates": 0.9523790950164666, "ibn_sina": 0.006807593134540645, "hypatia": 0.005880892147567166, "einstein": 0.012597336917451641, "sagan": 0.007315250843093265, "darwin": 0.015019831940880793}], "prompt_tokens": 615, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.012286566918535644, "einstein": 0.018454644591876835, "hypatia": 0.004471421570093671, "ibn_sina": 0.006220106501239739, "sagan": 0.010571965068651955, "socrates": 0.947995295349602}, "confidence": 0.8792284761037509}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.007286653720488884, "hypatia": 0.005400448709889409, "sagan": 0.008391780119708158, "socrates": 0.9480445169247135, "darwin": 0.014867451932552466, "einstein": 0.016009148592647675}, "confidence": 0.8793021066147155}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.0059405161975442, "darwin": 0.007050619482146086, "sagan": 0.00510210516641757, "ibn_sina": 0.007585503649862217, "einstein": 0.024245813678909663, "socrates": 0.9500754418251203}, "confidence": 0.8840797323408118}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.0058129823172400775, "einstein": 0.01253043283277484, "darwin": 0.006281895951539681, "sagan": 0.0096534954257128, "hypatia": 0.007214918136148576, "socrates": 0.958506275336584}, "confidence": 0.9029317475252799}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.02072378515744136, "sagan": 0.008859452034017536, "socrates": 0.9450135521359458, "ibn_sina": 0.009422828560209647, "darwin": 0.008729428827686823, "hypatia": 0.007250953284698772}, "confidence": 0.872531377397755}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.9523790950164666, "ibn_sina": 0.006807593134540645, "hypatia": 0.005880892147567166, "einstein": 0.012597336917451641, "sagan": 0.007315250843093265, "darwin": 0.015019831940880793}, "confidence": 0.8890536039819544}}, "usage": {"input_tokens": 615, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:34Z"}
{"id": "c-socrates-01-starter-03", "uid": "c107", "task": "kperm", "label": "socrates", "mode": "lev-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "prompt_sha256_by_order": ["0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296"], "choice": "socrates", "probs": {"darwin": 0.004965113482802749, "einstein": 0.0454609619946935, "hypatia": 0.004941172510433987, "ibn_sina": 0.05162345194000964, "sagan": 0.008547070794016532, "socrates": 0.8844622292780435}, "confidence": 0.7445526817984056, "seconds": 0.2191114468341766, "seconds_all_orders": 1.3146686810050596, "seconds_by_order": [0.21865897000134282, 0.21861955100030173, 0.21962088300097093, 0.2196964120012126, 0.21952185299960547, 0.21855101200162608], "argmax_stable": true, "spread": {"darwin": 0.0015938769331248397, "einstein": 0.04774471423439338, "hypatia": 0.0033231218394170793, "ibn_sina": 0.09528190098239897, "sagan": 0.00349652791450851, "socrates": 0.11424687350089302}, "spread_2dp": {"darwin": 0.0, "einstein": 0.05, "hypatia": 0.01, "ibn_sina": 0.09, "sagan": 0.0, "socrates": 0.12}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.004965113482802749, "einstein": 0.0454609619946935, "hypatia": 0.004941172510433987, "ibn_sina": 0.05162345194000964, "sagan": 0.008547070794016532, "socrates": 0.8844622292780435}, {"ibn_sina": 0.12462585371075863, "hypatia": 0.00598081365088903, "sagan": 0.007545027231518529, "socrates": 0.8114562051187206, "darwin": 0.004617928165639736, "einstein": 0.04577417212247355}, {"hypatia": 0.008264294349851067, "darwin": 0.0035726484652063733, "sagan": 0.005820270044331957, "ibn_sina": 0.02934395272835966, "einstein": 0.02729575579263733, "socrates": 0.9257030786196137}, {"ibn_sina": 0.10806045338156389, "einstein": 0.028260863593981392, "darwin": 0.004005204107028953, "sagan": 0.007298107030334622, "hypatia": 0.0075749214225920136, "socrates": 0.8448004504644991}, {"einstein": 0.0710377548734516, "sagan": 0.009316797958840467, "socrates": 0.8221292242812825, "ibn_sina": 0.0865113787666598, "darwin": 0.0033712365496779096, "hypatia": 0.007633607570087631}, {"socrates": 0.9103995338705495, "ibn_sina": 0.04990367572811627, "hypatia": 0.006750819526982663, "einstein": 0.023293040639058217, "sagan": 0.006184424109725485, "darwin": 0.003468506125567798}], "prompt_tokens": 561, "eval_tokens": 0, "lev_responses": [{"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"darwin": 0.004965113482802749, "einstein": 0.0454609619946935, "hypatia": 0.004941172510433987, "ibn_sina": 0.05162345194000964, "sagan": 0.008547070794016532, "socrates": 0.8844622292780435}, "confidence": 0.7445526817984056}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.12462585371075863, "hypatia": 0.00598081365088903, "sagan": 0.007545027231518529, "socrates": 0.8114562051187206, "darwin": 0.004617928165639736, "einstein": 0.04577417212247355}, "confidence": 0.6114424886808429}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"hypatia": 0.008264294349851067, "darwin": 0.0035726484652063733, "sagan": 0.005820270044331957, "ibn_sina": 0.02934395272835966, "einstein": 0.02729575579263733, "socrates": 0.9257030786196137}, "confidence": 0.8303767042400703}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"ibn_sina": 0.10806045338156389, "einstein": 0.028260863593981392, "darwin": 0.004005204107028953, "sagan": 0.007298107030334622, "hypatia": 0.0075749214225920136, "socrates": 0.8448004504644991}, "confidence": 0.6715482670742212}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"einstein": 0.0710377548734516, "sagan": 0.009316797958840467, "socrates": 0.8221292242812825, "ibn_sina": 0.0865113787666598, "darwin": 0.0033712365496779096, "hypatia": 0.007633607570087631}, "confidence": 0.6263001791386316}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}, {"model": "Qwen/Qwen3.5-4B", "answers": {"q": {"type": "choice", "choice": "socrates", "probabilities": {"socrates": 0.9103995338705495, "ibn_sina": 0.04990367572811627, "hypatia": 0.006750819526982663, "einstein": 0.023293040639058217, "sagan": 0.006184424109725485, "darwin": 0.003468506125567798}, "confidence": 0.7983473260788418}}, "usage": {"input_tokens": 561, "output_tokens": 0, "cached_input_tokens": 81}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "runtime": "lev serve 0.1.1 (git 0ea955f5aee0) over HTTP on 127.0.0.1, one question per request", "lev_revision": "f8ef71157ec06a7d3b6435bc0756f9d735c33748", "base_revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", "lev_commit": "0ea955f5aee0d56d906121d1779af20fa8b922ed", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.17.0", "peft": "0.21.0", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention)", "arm": "lev-4b", "model": "interfaze-ai/lev@f8ef7115 on Qwen/Qwen3.5-4B@851bf6e8", "quant": "bf16 backbone + LoRA r=32 (PEFT, unmerged) + mode-B head fp32; calibrated as served; order_average on", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 0, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T13:19:36Z"}