census: longest page 82500 tokens, num_ctx 131072, fits=True ANNOUNCE box benchbox · card GPU-46890836 · arm jev-readout · task a · start 2026-09-21T17:00:54Z on-01 yes -> yes 2.99s on-02 yes -> yes 3.50s on-03 yes -> yes 2.59s on-04 yes -> yes 2.83s on-05 yes -> yes 5.38s on-06 yes -> yes 5.80s on-07 yes -> yes 3.21s on-08 yes -> yes 2.36s on-09 yes -> yes 3.47s on-10 yes -> yes 4.69s on-11 yes -> yes 3.92s on-12 yes -> yes 1.97s on-13 yes -> yes 3.53s on-14 yes -> yes 12.18s on-15 yes -> yes 3.36s on-16 yes -> yes 2.71s on-17 yes -> yes 4.37s on-18 yes -> yes 3.11s on-19 yes -> yes 3.12s on-20 yes -> yes 3.05s on-21 yes -> yes 3.00s on-22 yes -> yes 3.62s on-23 yes -> yes 3.93s on-24 yes -> yes 3.71s on-25 yes -> yes 3.11s on-26 yes -> yes 3.13s on-27 yes -> yes 5.90s on-28 yes -> yes 5.12s on-29 yes -> yes 28.26s on-30 yes -> yes 3.74s on-31 yes -> yes 2.85s on-32 yes -> yes 3.15s on-33 yes -> yes 11.52s on-34 yes -> yes 6.99s on-35 yes -> yes 4.59s on-36 yes -> yes 4.87s on-37 yes -> yes 3.54s on-38 yes -> yes 4.52s on-39 yes -> yes 3.98s on-40 yes -> yes 7.25s on-41 yes -> yes 4.42s on-42 yes -> yes 3.41s off-01 no -> no 3.56s off-02 no -> no 2.82s off-03 no -> no 5.02s off-04 no -> no 39.81s off-05 no -> no 5.87s off-06 no -> no 3.10s off-07 no -> no 2.74s off-08 no -> no 3.14s off-09 no -> no 3.02s off-10 no -> no 2.30s off-11 no -> no 4.41s off-12 no -> no 3.10s off-13 no -> no 2.99s off-14 no -> no 28.02s off-15 no -> no 6.84s off-16 no -> no 3.07s off-17 no -> no 1.96s off-18 no -> no 12.08s off-19 no -> no 4.05s off-20 no -> no 3.45s off-21 no -> no 3.41s {"n": 63, "n_labelled": 63, "correct": 63, "errors": 0, "brier": 0.0037941824547977916, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": ANNOUNCE box benchbox · card GPU-46890836 · arm jev-readout · task b · start 2026-09-21T17:06:52Z on-01 - -> the-bill 40.05s on-02 - -> head-to-head 11.69s on-03 - -> for-the-technically-curious 2.63s on-04 - -> is-this-ai 2.96s on-05 - -> forty-one-releases-licence-by-lice 5.43s on-06 - -> how-to-listen 5.91s on-07 - -> the-bill 3.37s on-08 - -> five-minutes-of-training-honestly- 2.60s on-09 - -> two-local-models-sixteen-days-earl 3.29s on-10 - -> two-hours-at-12-tok-s-the-test-tha 4.90s on-11 - -> what-the-battery-was-for-the-instr 4.14s on-12 - -> top 2.01s on-13 - -> the-answer-in-one-paragraph 3.53s on-14 - -> the-sound-is-made-on-your-device 12.21s on-15 - -> the-developer 3.41s on-16 - -> the-table-that-answers-the-questio 2.93s on-17 - -> finding-one-a-picture-is-199-token 4.57s on-18 - -> finding-four-the-argument-with-our 3.13s on-19 - -> the-world-is-a-file 3.15s on-20 - -> two-kinds-of-there-all-along 3.05s on-21 - -> the-punchline 3.00s on-22 - -> the-day-the-box-got-faster 3.65s on-23 - -> the-same-question-down-the-whole-l 4.06s on-24 - -> the-part-nobody-warns-you-about-th 3.81s on-25 - -> a-measurement-from-our-own-card 3.09s on-26 - -> the-nerd-shelf-the-numbers-from-th 3.21s on-27 - -> the-seat-question 6.08s on-28 - -> where-the-file-came-from 5.43s on-29 - -> 2-2826-over-56-metered-calls-again 28.39s on-30 - -> the-first-promise-your-visit-stays 3.73s on-31 - -> the-thorough-walk 3.00s on-32 - -> receipts 3.15s on-33 - -> top 11.65s on-34 - -> top 7.09s on-35 - -> vendors 4.77s on-36 - -> what-to-take-with-you 4.89s on-37 - -> the-instrument 3.73s on-38 - -> provenance-4cae43 4.72s on-39 - -> how-to-read-this-page 4.02s on-40 - -> the-blind-sheet-unsealed 7.35s on-41 - -> the-blind-sitting 4.53s on-42 ['no template tail — no closing sentence of the shape "the article does not specify what was left out, moved or not measured" where the article carries no such statement', "an operator, never the definite one — the hub's voice law (an operator 2026-09-03, re-stated 2026-09-04)", "the prediction's own subject kept — the page predicts that a few COMPANIES will sell intelligence as a utility; the commodity sentence is about SOLD INTELLIGENCE and may not be re-subjected onto the prediction"] -> the-indifferent-prediction 3.50s {"n": 42, "n_labelled": 1, "correct": 0, "errors": 0, "brier": 0.946055144701946, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, ANNOUNCE box benchbox · card GPU-46890836 · arm jev-readout · task c · start 2026-09-21T17:11:15Z c-darwin-01-dessert-01 darwin -> darwin 1.46s c-darwin-01-dessert-01 darwin -> darwin 0.96s c-darwin-01-dessert-01 darwin -> darwin 1.02s c-darwin-01-dessert-02 darwin -> darwin 1.05s c-darwin-01-dessert-03 darwin -> darwin 1.02s c-darwin-01-dessert-03 darwin -> darwin 1.06s c-darwin-01-main-01 darwin -> darwin 1.02s c-darwin-01-main-02 darwin -> ibn_sina 1.10s c-darwin-01-main-02 darwin -> darwin 0.94s c-darwin-01-main-02 darwin -> darwin 1.00s c-darwin-01-main-02 darwin -> darwin 1.01s c-darwin-01-main-03 darwin -> darwin 0.99s c-darwin-01-starter-01 darwin -> darwin 1.00s c-darwin-01-starter-01 darwin -> darwin 0.98s c-darwin-01-starter-01 darwin -> darwin 0.97s c-darwin-01-starter-02 darwin -> darwin 1.10s c-darwin-01-starter-03 darwin -> darwin 1.09s c-darwin-01-starter-03 darwin -> darwin 1.07s c-einstein-01-dessert-01 einstein -> sagan 1.02s c-einstein-01-dessert-01 einstein -> sagan 1.04s c-einstein-01-dessert-02 einstein -> einstein 1.01s c-einstein-01-dessert-02 einstein -> einstein 1.06s c-einstein-01-dessert-03 einstein -> einstein 0.93s c-einstein-01-main-01 einstein -> hypatia 1.09s c-einstein-01-main-01 einstein -> einstein 1.05s c-einstein-01-main-02 einstein -> einstein 1.05s c-einstein-01-main-03 einstein -> sagan 1.03s c-einstein-01-main-03 einstein -> einstein 1.06s c-einstein-01-main-03 einstein -> einstein 1.00s c-einstein-01-main-03 einstein -> einstein 1.04s c-einstein-01-starter-01 einstein -> einstein 1.06s c-einstein-01-starter-02 einstein -> einstein 1.07s c-einstein-01-starter-02 einstein -> darwin 1.02s c-einstein-01-starter-03 einstein -> einstein 1.02s c-einstein-01-starter-03 einstein -> einstein 1.02s c-einstein-01-starter-03 einstein -> einstein 1.00s c-hypatia-01-dessert-01 hypatia -> hypatia 1.05s c-hypatia-01-dessert-02 hypatia -> hypatia 0.99s c-hypatia-01-dessert-02 hypatia -> sagan 1.03s c-hypatia-01-dessert-03 hypatia -> ibn_sina 0.95s c-hypatia-01-main-01 hypatia -> hypatia 1.02s c-hypatia-01-main-01 hypatia -> hypatia 1.02s c-hypatia-01-main-01 hypatia -> hypatia 1.08s c-hypatia-01-main-01 hypatia -> hypatia 1.02s c-hypatia-01-main-01 hypatia -> sagan 0.99s c-hypatia-01-main-02 hypatia -> ibn_sina 0.99s c-hypatia-01-main-03 hypatia -> ibn_sina 1.02s c-hypatia-01-main-03 hypatia -> hypatia 1.09s c-hypatia-01-main-03 hypatia -> hypatia 1.02s c-hypatia-01-starter-01 hypatia -> hypatia 1.01s c-hypatia-01-starter-01 hypatia -> darwin 0.95s c-hypatia-01-starter-02 hypatia -> hypatia 0.97s c-hypatia-01-starter-02 hypatia -> hypatia 0.99s c-hypatia-01-starter-02 hypatia -> hypatia 0.96s c-ibn_sina-01-dessert-01 ibn_sina -> ibn_sina 0.98s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 1.07s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 0.96s c-ibn_sina-01-dessert-03 ibn_sina -> ibn_sina 0.92s c-ibn_sina-01-dessert-03 ibn_sina -> ibn_sina 1.03s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 0.98s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.07s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.03s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.09s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 0.99s c-ibn_sina-01-starter-01 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-starter-02 ibn_sina -> ibn_sina 0.95s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 0.94s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.07s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 0.98s c-sagan-01-dessert-01 sagan -> sagan 1.13s c-sagan-01-dessert-01 sagan -> sagan 0.99s c-sagan-01-dessert-01 sagan -> sagan 1.03s c-sagan-01-dessert-03 sagan -> sagan 1.00s c-sagan-01-dessert-03 sagan -> sagan 0.95s c-sagan-01-dessert-03 sagan -> sagan 1.04s c-sagan-01-dessert-03 sagan -> sagan 0.98s c-sagan-01-dessert-03 sagan -> sagan 1.03s c-sagan-01-main-01 sagan -> sagan 1.00s c-sagan-01-main-02 sagan -> sagan 1.04s c-sagan-01-main-02 sagan -> sagan 0.91s c-sagan-01-main-03 sagan -> sagan 0.98s c-sagan-01-main-03 sagan -> sagan 0.97s c-sagan-01-starter-01 sagan -> sagan 1.07s c-sagan-01-starter-02 sagan -> sagan 1.02s c-sagan-01-starter-02 sagan -> sagan 1.08s c-sagan-01-starter-03 sagan -> darwin 1.03s c-sagan-01-starter-03 sagan -> einstein 1.03s c-socrates-01-dessert-01 socrates -> socrates 1.09s c-socrates-01-dessert-01 socrates -> socrates 1.02s c-socrates-01-dessert-01 socrates -> socrates 1.00s c-socrates-01-dessert-02 socrates -> sagan 1.10s c-socrates-01-dessert-02 socrates -> socrates 0.98s c-socrates-01-dessert-02 socrates -> socrates 1.01s c-socrates-01-main-01 socrates -> socrates 0.96s c-socrates-01-main-02 socrates -> socrates 1.04s c-socrates-01-main-02 socrates -> ibn_sina 0.97s c-socrates-01-main-02 socrates -> socrates 1.03s c-socrates-01-main-02 socrates -> ibn_sina 0.97s c-socrates-01-main-03 socrates -> sagan 0.99s c-socrates-01-main-03 socrates -> socrates 1.02s c-socrates-01-starter-01 socrates -> socrates 1.05s c-socrates-01-starter-02 socrates -> socrates 1.02s c-socrates-01-starter-03 socrates -> socrates 1.04s c-socrates-01-starter-03 socrates -> socrates 0.98s c-socrates-01-starter-03 socrates -> socrates 1.07s {"n": 108, "n_labelled": 108, "correct": 90, "errors": 0, "brier": 0.2874528932055017, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 3, "sum_p": 0.8672172052495464, "correct": 1, "mean_p": 0.2890724017498488, "accuracy": 0.3333333333333333, "gap": -0.04426093158348454}, {"lo": 0.3, "hi": 0.4, "n": 11, "sum_p": 3.929440965782504, "correct": 5, "mean_p": 0.3572219059802276, "accuracy": 0 census: reusing rows/census.hf.co_mradermacher_jevify-gemma4-26b-a4b-GGUF-Q4_K_M.json census: longest page 82500 tokens, num_ctx 131072, fits=True ANNOUNCE box benchbox · card GPU-46890836 · arm jev-generate · task a · start 2026-09-21T17:13:16Z on-01 yes -> yes 39.21s on-02 yes -> yes 11.60s on-03 yes -> yes 2.63s on-04 yes -> yes 2.85s on-05 yes -> yes 5.45s on-06 yes -> yes 5.88s on-07 yes -> yes 3.18s on-08 yes -> yes 2.46s on-09 yes -> yes 3.36s on-10 yes -> yes 4.82s on-11 yes -> yes 3.95s on-12 yes -> yes 2.08s on-13 yes -> yes 3.47s on-14 yes -> yes 12.21s on-15 yes -> yes 3.48s on-16 yes -> yes 2.75s on-17 yes -> yes 4.37s on-18 yes -> yes 3.09s on-19 yes -> yes 3.21s on-20 yes -> yes 3.06s on-21 yes -> yes 3.03s on-22 yes -> yes 3.68s on-23 yes -> yes 3.98s on-24 yes -> yes 3.69s on-25 yes -> yes 3.13s on-26 yes -> yes 3.09s on-27 yes -> yes 6.00s on-28 yes -> yes 5.24s on-29 yes -> yes 28.21s on-30 yes -> yes 3.84s on-31 yes -> yes 2.81s on-32 yes -> yes 3.17s on-33 yes -> yes 11.53s on-34 yes -> yes 6.95s on-35 yes -> yes 4.62s on-36 yes -> yes 4.86s on-37 yes -> yes 3.55s on-38 yes -> yes 4.55s on-39 yes -> yes 3.88s on-40 yes -> yes 7.30s on-41 yes -> yes 4.42s on-42 yes -> yes 3.41s off-01 no -> no 3.62s off-02 no -> no 2.83s off-03 no -> no 5.14s off-04 no -> no 39.82s off-05 no -> no 6.06s off-06 no -> no 3.07s off-07 no -> no 2.86s off-08 no -> no 3.25s off-09 no -> no 3.06s off-10 no -> no 2.30s off-11 no -> no 4.37s off-12 no -> no 3.10s off-13 no -> no 3.09s off-14 no -> no 28.09s off-15 no -> no 6.85s off-16 no -> no 3.01s off-17 no -> no 2.05s off-18 no -> no 12.20s off-19 no -> no 4.10s off-20 no -> no 3.42s off-21 no -> no 3.46s {"n": 63, "n_labelled": 63, "correct": 63, "errors": 0, "brier": 8.349285404259994e-05, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": ANNOUNCE box benchbox · card GPU-46890836 · arm jev-generate · task b · start 2026-09-21T17:20:00Z on-01 - -> the-bill 40.06s on-02 - -> head-to-head 11.62s on-03 - -> for-the-technically-curious 2.64s on-04 - -> is-this-ai 2.92s on-05 - -> forty-one-releases-licence-by-lice 5.47s on-06 - -> how-to-listen 5.97s on-07 - -> the-bill 3.37s on-08 - -> five-minutes-of-training-honestly- 2.62s on-09 - -> two-local-models-sixteen-days-earl 3.34s on-10 - -> two-hours-at-12-tok-s-the-test-tha 4.87s on-11 - -> what-the-battery-was-for-the-instr 4.07s on-12 - -> top 1.99s on-13 - -> the-answer-in-one-paragraph 3.59s on-14 - -> the-sound-is-made-on-your-device 12.21s on-15 - -> the-developer 3.48s on-16 - -> the-table-that-answers-the-questio 2.85s on-17 - -> finding-one-a-picture-is-199-token 4.63s on-18 - -> finding-four-the-argument-with-our 3.11s on-19 - -> the-world-is-a-file 3.24s on-20 - -> two-kinds-of-there-all-along 3.03s on-21 - -> the-punchline 3.05s on-22 - -> the-day-the-box-got-faster 3.60s on-23 - -> the-same-question-down-the-whole-l 4.02s on-24 - -> the-part-nobody-warns-you-about-th 3.86s on-25 - -> a-measurement-from-our-own-card 3.15s on-26 - -> the-nerd-shelf-the-numbers-from-th 3.12s on-27 - -> the-seat-question 6.08s on-28 - -> where-the-file-came-from 5.44s on-29 - -> 2-2826-over-56-metered-calls-again 28.45s on-30 - -> the-first-promise-your-visit-stays 3.83s on-31 - -> the-thorough-walk 3.02s on-32 - -> receipts 3.13s on-33 - -> top 11.64s on-34 - -> top 7.05s on-35 - -> vendors 4.74s on-36 - -> what-to-take-with-you 4.89s on-37 - -> the-instrument 3.70s on-38 - -> provenance-4cae43 4.72s on-39 - -> how-to-read-this-page 4.02s on-40 - -> the-blind-sheet-unsealed 7.44s on-41 - -> the-blind-sitting 4.48s on-42 ['no template tail — no closing sentence of the shape "the article does not specify what was left out, moved or not measured" where the article carries no such statement', "an operator, never the definite one — the hub's voice law (an operator 2026-09-03, re-stated 2026-09-04)", "the prediction's own subject kept — the page predicts that a few COMPANIES will sell intelligence as a utility; the commodity sentence is about SOLD INTELLIGENCE and may not be re-subjected onto the prediction"] -> the-indifferent-prediction 3.55s {"n": 42, "n_labelled": 1, "correct": 0, "errors": 0, "brier": 1.0, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": n ANNOUNCE box benchbox · card GPU-46890836 · arm jev-generate · task c · start 2026-09-21T17:24:23Z c-darwin-01-dessert-01 darwin -> darwin 1.45s c-darwin-01-dessert-01 darwin -> darwin 1.02s c-darwin-01-dessert-01 darwin -> darwin 1.02s c-darwin-01-dessert-02 darwin -> darwin 1.04s c-darwin-01-dessert-03 darwin -> darwin 1.05s c-darwin-01-dessert-03 darwin -> darwin 1.07s c-darwin-01-main-01 darwin -> darwin 1.05s c-darwin-01-main-02 darwin -> ibn_sina 1.03s c-darwin-01-main-02 darwin -> darwin 1.06s c-darwin-01-main-02 darwin -> darwin 1.04s c-darwin-01-main-02 darwin -> darwin 1.06s c-darwin-01-main-03 darwin -> darwin 1.06s c-darwin-01-starter-01 darwin -> darwin 0.94s c-darwin-01-starter-01 darwin -> darwin 0.99s c-darwin-01-starter-01 darwin -> darwin 1.01s c-darwin-01-starter-02 darwin -> darwin 1.10s c-darwin-01-starter-03 darwin -> darwin 1.00s c-darwin-01-starter-03 darwin -> darwin 1.06s c-einstein-01-dessert-01 einstein -> sagan 1.03s c-einstein-01-dessert-01 einstein -> sagan 1.04s c-einstein-01-dessert-02 einstein -> einstein 1.03s c-einstein-01-dessert-02 einstein -> einstein 1.01s c-einstein-01-dessert-03 einstein -> einstein 0.96s c-einstein-01-main-01 einstein -> hypatia 1.01s c-einstein-01-main-01 einstein -> einstein 1.10s c-einstein-01-main-02 einstein -> einstein 1.03s c-einstein-01-main-03 einstein -> sagan 1.05s c-einstein-01-main-03 einstein -> einstein 1.07s c-einstein-01-main-03 einstein -> einstein 1.06s c-einstein-01-main-03 einstein -> einstein 1.00s c-einstein-01-starter-01 einstein -> einstein 1.05s c-einstein-01-starter-02 einstein -> einstein 0.97s c-einstein-01-starter-02 einstein -> darwin 0.95s c-einstein-01-starter-03 einstein -> einstein 1.02s c-einstein-01-starter-03 einstein -> einstein 1.06s c-einstein-01-starter-03 einstein -> einstein 0.95s c-hypatia-01-dessert-01 hypatia -> hypatia 1.03s c-hypatia-01-dessert-02 hypatia -> hypatia 1.02s c-hypatia-01-dessert-02 hypatia -> sagan 1.00s c-hypatia-01-dessert-03 hypatia -> ibn_sina 0.97s c-hypatia-01-main-01 hypatia -> hypatia 1.07s c-hypatia-01-main-01 hypatia -> hypatia 1.05s c-hypatia-01-main-01 hypatia -> hypatia 1.05s c-hypatia-01-main-01 hypatia -> hypatia 0.95s c-hypatia-01-main-01 hypatia -> sagan 1.01s c-hypatia-01-main-02 hypatia -> ibn_sina 1.01s c-hypatia-01-main-03 hypatia -> ibn_sina 1.08s c-hypatia-01-main-03 hypatia -> hypatia 1.04s c-hypatia-01-main-03 hypatia -> hypatia 1.03s c-hypatia-01-starter-01 hypatia -> hypatia 0.99s c-hypatia-01-starter-01 hypatia -> darwin 1.00s c-hypatia-01-starter-02 hypatia -> hypatia 1.08s c-hypatia-01-starter-02 hypatia -> hypatia 1.05s c-hypatia-01-starter-02 hypatia -> hypatia 0.97s c-ibn_sina-01-dessert-01 ibn_sina -> ibn_sina 0.99s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 1.06s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 1.10s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 0.98s c-ibn_sina-01-dessert-03 ibn_sina -> ibn_sina 0.96s c-ibn_sina-01-dessert-03 ibn_sina -> ibn_sina 1.04s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.09s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.06s c-ibn_sina-01-starter-01 ibn_sina -> ibn_sina 1.06s c-ibn_sina-01-starter-02 ibn_sina -> ibn_sina 0.93s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 0.99s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.09s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.07s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.02s c-sagan-01-dessert-01 sagan -> sagan 1.01s c-sagan-01-dessert-01 sagan -> sagan 0.98s c-sagan-01-dessert-01 sagan -> sagan 0.96s c-sagan-01-dessert-03 sagan -> sagan 1.04s c-sagan-01-dessert-03 sagan -> sagan 1.01s c-sagan-01-dessert-03 sagan -> sagan 1.06s c-sagan-01-dessert-03 sagan -> sagan 1.03s c-sagan-01-dessert-03 sagan -> sagan 0.95s c-sagan-01-main-01 sagan -> sagan 1.03s c-sagan-01-main-02 sagan -> sagan 0.97s c-sagan-01-main-02 sagan -> sagan 1.07s c-sagan-01-main-03 sagan -> sagan 1.00s c-sagan-01-main-03 sagan -> sagan 1.09s c-sagan-01-starter-01 sagan -> sagan 0.99s c-sagan-01-starter-02 sagan -> sagan 1.02s c-sagan-01-starter-02 sagan -> sagan 1.06s c-sagan-01-starter-03 sagan -> darwin 0.93s c-sagan-01-starter-03 sagan -> einstein 1.04s c-socrates-01-dessert-01 socrates -> socrates 1.06s c-socrates-01-dessert-01 socrates -> socrates 1.07s c-socrates-01-dessert-01 socrates -> socrates 1.03s c-socrates-01-dessert-02 socrates -> sagan 1.03s c-socrates-01-dessert-02 socrates -> socrates 0.98s c-socrates-01-dessert-02 socrates -> socrates 1.04s c-socrates-01-main-01 socrates -> socrates 0.96s c-socrates-01-main-02 socrates -> socrates 1.04s c-socrates-01-main-02 socrates -> ibn_sina 1.05s c-socrates-01-main-02 socrates -> socrates 1.02s c-socrates-01-main-02 socrates -> ibn_sina 1.04s c-socrates-01-main-03 socrates -> sagan 1.03s c-socrates-01-main-03 socrates -> socrates 1.03s c-socrates-01-starter-01 socrates -> socrates 0.95s c-socrates-01-starter-02 socrates -> socrates 1.01s c-socrates-01-starter-03 socrates -> socrates 1.08s c-socrates-01-starter-03 socrates -> socrates 1.05s c-socrates-01-starter-03 socrates -> socrates 1.09s {"n": 108, "n_labelled": 108, "correct": 90, "errors": 0, "brier": 0.3333333333333333, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": census: longest page 82500 tokens, num_ctx 131072, fits=True ANNOUNCE box benchbox · card GPU-46890836 · arm base-readout · task a · start 2026-09-21T17:31:21Z on-01 yes -> yes 3.49s on-02 yes -> yes 3.63s on-03 yes -> yes 2.82s on-04 yes -> yes 3.03s on-05 yes -> yes 5.71s on-06 yes -> yes 6.19s on-07 yes -> yes 3.44s on-08 yes -> yes 2.68s on-09 yes -> yes 3.42s on-10 yes -> yes 5.19s on-11 yes -> yes 4.24s on-12 yes -> yes 2.17s on-13 yes -> yes 3.75s on-14 yes -> yes 12.94s on-15 yes -> yes 3.68s on-16 yes -> yes 3.01s on-17 yes -> yes 4.78s on-18 yes -> yes 3.24s on-19 yes -> yes 3.32s on-20 yes -> yes 3.17s on-21 yes -> yes 3.17s on-22 yes -> yes 3.85s on-23 yes -> yes 4.29s on-24 yes -> yes 3.99s on-25 yes -> yes 3.28s on-26 yes -> yes 3.30s on-27 yes -> yes 6.36s on-28 yes -> yes 5.66s on-29 yes -> yes 29.96s on-30 yes -> yes 3.90s on-31 yes -> yes 3.03s on-32 yes -> yes 3.36s on-33 yes -> yes 12.33s on-34 yes -> yes 7.37s on-35 yes -> yes 4.95s on-36 yes -> yes 5.10s on-37 yes -> yes 3.81s on-38 yes -> yes 4.77s on-39 yes -> yes 4.21s on-40 yes -> yes 7.61s on-41 yes -> yes 4.65s on-42 yes -> yes 3.69s off-01 no -> no 3.81s off-02 no -> no 3.04s off-03 no -> no 5.46s off-04 no -> no 41.94s off-05 no -> no 6.30s off-06 no -> no 3.31s off-07 no -> no 2.96s off-08 no -> no 3.36s off-09 no -> no 3.09s off-10 no -> no 2.50s off-11 no -> no 4.70s off-12 no -> no 3.31s off-13 no -> no 3.22s off-14 no -> no 29.87s off-15 no -> no 7.26s off-16 no -> no 3.18s off-17 no -> no 2.14s off-18 no -> no 12.90s off-19 no -> no 4.29s off-20 no -> no 3.63s off-21 no -> no 3.70s {"n": 63, "n_labelled": 63, "correct": 63, "errors": 0, "brier": 8.38398270530722e-05, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": ANNOUNCE box benchbox · card GPU-46890836 · arm base-readout · task b · start 2026-09-21T17:37:43Z on-01 - -> the-bill 42.41s on-02 - -> head-to-head 12.42s on-03 - -> for-the-technically-curious 2.85s on-04 - -> is-this-ai 3.18s on-05 - -> top 5.65s on-06 - -> how-to-listen 6.20s on-07 - -> the-bill 3.48s on-08 - -> five-minutes-of-training-honestly- 2.82s on-09 - -> sixteen-of-nineteen 3.55s on-10 - -> two-hours-at-12-tok-s-the-test-tha 5.24s on-11 - -> what-the-battery-was-for-the-instr 4.42s on-12 - -> top 2.15s on-13 - -> the-answer-in-one-paragraph 3.85s on-14 - -> what-happens-when-you-ask-the-geni 12.97s on-15 - -> the-developer 3.67s on-16 - -> the-table-that-answers-the-questio 3.05s on-17 - -> finding-one-a-picture-is-199-token 4.92s on-18 - -> finding-four-the-argument-with-our 3.32s on-19 - -> the-world-is-a-file 3.42s on-20 - -> two-kinds-of-there-all-along 3.29s on-21 - -> what-we-can-show-you 3.22s on-22 - -> the-day-the-box-got-faster 3.94s on-23 - -> the-same-question-down-the-whole-l 4.33s on-24 - -> the-payroll-and-the-meeting 4.02s on-25 - -> a-measurement-from-our-own-card 3.37s on-26 - -> the-third-librarian-never-left-you 3.34s on-27 - -> what-to-take-with-you 6.55s on-28 - -> where-the-file-came-from 5.68s on-29 - -> top 30.13s on-30 - -> the-first-promise-your-visit-stays 4.00s on-31 - -> the-thorough-walk 3.18s on-32 - -> receipts 3.33s on-33 - -> top 12.38s on-34 - -> top 7.44s on-35 - -> vendors 4.98s on-36 - -> top 5.12s on-37 - -> the-instrument 3.89s on-38 - -> provenance-4cae43 4.99s on-39 - -> top 4.34s on-40 - -> the-blind-sheet-unsealed 7.88s on-41 - -> the-blind-sitting 4.81s on-42 ['no template tail — no closing sentence of the shape "the article does not specify what was left out, moved or not measured" where the article carries no such statement', "an operator, never the definite one — the hub's voice law (an operator 2026-09-03, re-stated 2026-09-04)", "the prediction's own subject kept — the page predicts that a few COMPANIES will sell intelligence as a utility; the commodity sentence is about SOLD INTELLIGENCE and may not be re-subjected onto the prediction"] -> what-an-operator-saw 3.74s {"n": 42, "n_labelled": 1, "correct": 0, "errors": 0, "brier": 0.9999999016716582, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null ANNOUNCE box benchbox · card GPU-46890836 · arm base-readout · task c · start 2026-09-21T17:42:21Z c-darwin-01-dessert-01 darwin -> darwin 1.54s c-darwin-01-dessert-01 darwin -> darwin 0.99s c-darwin-01-dessert-01 darwin -> ibn_sina 1.00s c-darwin-01-dessert-02 darwin -> darwin 1.07s c-darwin-01-dessert-03 darwin -> hypatia 1.11s c-darwin-01-dessert-03 darwin -> darwin 1.10s c-darwin-01-main-01 darwin -> darwin 0.99s c-darwin-01-main-02 darwin -> ibn_sina 1.04s c-darwin-01-main-02 darwin -> darwin 1.06s c-darwin-01-main-02 darwin -> darwin 1.03s c-darwin-01-main-02 darwin -> darwin 1.09s c-darwin-01-main-03 darwin -> darwin 1.03s c-darwin-01-starter-01 darwin -> sagan 0.99s c-darwin-01-starter-01 darwin -> darwin 0.97s c-darwin-01-starter-01 darwin -> sagan 0.96s c-darwin-01-starter-02 darwin -> darwin 1.01s c-darwin-01-starter-03 darwin -> darwin 1.14s c-darwin-01-starter-03 darwin -> darwin 1.05s c-einstein-01-dessert-01 einstein -> sagan 1.00s c-einstein-01-dessert-01 einstein -> sagan 1.01s c-einstein-01-dessert-02 einstein -> einstein 1.09s c-einstein-01-dessert-02 einstein -> einstein 1.02s c-einstein-01-dessert-03 einstein -> einstein 1.09s c-einstein-01-main-01 einstein -> socrates 0.99s c-einstein-01-main-01 einstein -> einstein 1.12s c-einstein-01-main-02 einstein -> einstein 0.95s c-einstein-01-main-03 einstein -> sagan 1.03s c-einstein-01-main-03 einstein -> einstein 1.02s c-einstein-01-main-03 einstein -> einstein 1.00s c-einstein-01-main-03 einstein -> einstein 1.03s c-einstein-01-starter-01 einstein -> einstein 1.04s c-einstein-01-starter-02 einstein -> einstein 1.00s c-einstein-01-starter-02 einstein -> einstein 1.01s c-einstein-01-starter-03 einstein -> einstein 1.02s c-einstein-01-starter-03 einstein -> einstein 1.00s c-einstein-01-starter-03 einstein -> einstein 0.98s c-hypatia-01-dessert-01 hypatia -> hypatia 1.04s c-hypatia-01-dessert-02 hypatia -> hypatia 1.02s c-hypatia-01-dessert-02 hypatia -> sagan 1.10s c-hypatia-01-dessert-03 hypatia -> hypatia 0.97s c-hypatia-01-main-01 hypatia -> hypatia 1.04s c-hypatia-01-main-01 hypatia -> einstein 1.02s c-hypatia-01-main-01 hypatia -> hypatia 1.02s c-hypatia-01-main-01 hypatia -> hypatia 1.10s c-hypatia-01-main-01 hypatia -> sagan 1.03s c-hypatia-01-main-02 hypatia -> einstein 1.03s c-hypatia-01-main-03 hypatia -> einstein 0.98s c-hypatia-01-main-03 hypatia -> hypatia 1.03s c-hypatia-01-main-03 hypatia -> hypatia 1.15s c-hypatia-01-starter-01 hypatia -> hypatia 1.00s c-hypatia-01-starter-01 hypatia -> socrates 0.99s c-hypatia-01-starter-02 hypatia -> hypatia 1.01s c-hypatia-01-starter-02 hypatia -> hypatia 1.04s c-hypatia-01-starter-02 hypatia -> hypatia 0.99s c-ibn_sina-01-dessert-01 ibn_sina -> sagan 1.00s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 1.02s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 1.12s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 1.06s c-ibn_sina-01-dessert-03 ibn_sina -> ibn_sina 0.98s c-ibn_sina-01-dessert-03 ibn_sina -> ibn_sina 1.03s c-ibn_sina-01-main-02 ibn_sina -> hypatia 0.97s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.00s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.05s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.05s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.11s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-starter-01 ibn_sina -> hypatia 0.97s c-ibn_sina-01-starter-02 ibn_sina -> socrates 1.05s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 0.99s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.06s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.06s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.05s c-sagan-01-dessert-01 sagan -> sagan 1.14s c-sagan-01-dessert-01 sagan -> sagan 1.03s c-sagan-01-dessert-01 sagan -> sagan 1.04s c-sagan-01-dessert-03 sagan -> sagan 0.94s c-sagan-01-dessert-03 sagan -> sagan 1.05s c-sagan-01-dessert-03 sagan -> sagan 0.99s c-sagan-01-dessert-03 sagan -> sagan 0.94s c-sagan-01-dessert-03 sagan -> sagan 1.08s c-sagan-01-main-01 sagan -> sagan 1.02s c-sagan-01-main-02 sagan -> sagan 1.05s c-sagan-01-main-02 sagan -> sagan 1.05s c-sagan-01-main-03 sagan -> sagan 1.08s c-sagan-01-main-03 sagan -> sagan 1.07s c-sagan-01-starter-01 sagan -> sagan 1.02s c-sagan-01-starter-02 sagan -> sagan 1.08s c-sagan-01-starter-02 sagan -> sagan 1.03s c-sagan-01-starter-03 sagan -> darwin 1.01s c-sagan-01-starter-03 sagan -> sagan 1.01s c-socrates-01-dessert-01 socrates -> socrates 1.06s c-socrates-01-dessert-01 socrates -> socrates 1.06s c-socrates-01-dessert-01 socrates -> socrates 0.93s c-socrates-01-dessert-02 socrates -> sagan 1.11s c-socrates-01-dessert-02 socrates -> socrates 1.02s c-socrates-01-dessert-02 socrates -> socrates 1.04s c-socrates-01-main-01 socrates -> socrates 1.01s c-socrates-01-main-02 socrates -> socrates 0.95s c-socrates-01-main-02 socrates -> socrates 1.02s c-socrates-01-main-02 socrates -> socrates 1.02s c-socrates-01-main-02 socrates -> socrates 1.02s c-socrates-01-main-03 socrates -> socrates 1.05s c-socrates-01-main-03 socrates -> socrates 1.05s c-socrates-01-starter-01 socrates -> socrates 0.95s c-socrates-01-starter-02 socrates -> socrates 0.95s c-socrates-01-starter-03 socrates -> socrates 1.05s c-socrates-01-starter-03 socrates -> socrates 0.93s c-socrates-01-starter-03 socrates -> socrates 1.01s {"n": 108, "n_labelled": 108, "correct": 87, "errors": 0, "brier": 0.363485888024185, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 2, "sum_p": 0.8639594982007308, "correct": census: reusing rows/census.gemma4-26b.json census: longest page 82500 tokens, num_ctx 131072, fits=True ANNOUNCE box benchbox · card GPU-46890836 · arm base-generate · task a · start 2026-09-21T17:44:23Z on-01 yes -> yes 41.45s on-02 yes -> yes 12.27s on-03 yes -> yes 2.90s on-04 yes -> yes 2.94s on-05 yes -> yes 5.74s on-06 yes -> yes 6.19s on-07 yes -> yes 3.49s on-08 yes -> yes 2.62s on-09 yes -> yes 3.49s on-10 yes -> yes 5.22s on-11 yes -> yes 4.30s on-12 yes -> yes 2.14s on-13 yes -> yes 3.76s on-14 yes -> yes 12.89s on-15 yes -> yes 3.66s on-16 yes -> yes 2.98s on-17 yes -> yes 4.76s on-18 yes -> yes 3.19s on-19 yes -> yes 3.45s on-20 yes -> yes 3.22s on-21 yes -> yes 3.24s on-22 yes -> yes 3.86s on-23 yes -> yes 4.28s on-24 yes -> yes 3.98s on-25 yes -> yes 3.32s on-26 yes -> yes 3.27s on-27 yes -> yes 6.38s on-28 yes -> yes 5.58s on-29 yes -> yes 29.96s on-30 yes -> yes 4.06s on-31 yes -> yes 3.12s on-32 yes -> yes 3.43s on-33 yes -> yes 12.25s on-34 yes -> yes 7.43s on-35 yes -> yes 4.91s on-36 yes -> yes 5.19s on-37 yes -> yes 3.84s on-38 yes -> yes 4.86s on-39 yes -> yes 4.23s on-40 yes -> yes 7.62s on-41 yes -> yes 4.58s on-42 yes -> yes 3.71s off-01 no -> no 3.82s off-02 no -> no 3.09s off-03 no -> no 5.48s off-04 no -> no 42.08s off-05 no -> no 6.32s off-06 no -> no 3.35s off-07 no -> no 2.91s off-08 no -> no 3.43s off-09 no -> no 3.17s off-10 no -> no 2.51s off-11 no -> no 4.76s off-12 no -> no 3.31s off-13 no -> no 3.23s off-14 no -> no 29.90s off-15 no -> no 7.26s off-16 no -> no 3.26s off-17 no -> no 2.17s off-18 no -> no 12.91s off-19 no -> no 4.27s off-20 no -> no 3.68s off-21 no -> no 3.74s {"n": 63, "n_labelled": 63, "correct": 63, "errors": 0, "brier": 8.349285404259994e-05, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": ANNOUNCE box benchbox · card GPU-46890836 · arm base-generate · task b · start 2026-09-21T17:51:32Z on-01 - -> the-bill 42.44s on-02 - -> head-to-head 12.41s on-03 - -> for-the-technically-curious 2.91s on-04 - -> is-this-ai 3.23s on-05 - -> top 5.78s on-06 - -> how-to-listen 6.26s on-07 - -> the-bill 3.51s on-08 - -> five-minutes-of-training-honestly- 2.83s on-09 - -> sixteen-of-nineteen 3.55s on-10 - -> two-hours-at-12-tok-s-the-test-tha 5.27s on-11 - -> what-the-battery-was-for-the-instr 4.42s on-12 - -> top 2.18s on-13 - -> the-answer-in-one-paragraph 3.88s on-14 - -> what-happens-when-you-ask-the-geni 13.05s on-15 - -> the-developer 3.73s on-16 - -> the-table-that-answers-the-questio 3.04s on-17 - -> finding-one-a-picture-is-199-token 4.87s on-18 - -> finding-four-the-argument-with-our 3.34s on-19 - -> the-world-is-a-file 3.47s on-20 - -> two-kinds-of-there-all-along 3.32s on-21 - -> what-we-can-show-you 3.23s on-22 - -> the-day-the-box-got-faster 3.93s on-23 - -> the-same-question-down-the-whole-l 4.30s on-24 - -> the-payroll-and-the-meeting 4.12s on-25 - -> a-measurement-from-our-own-card 3.32s on-26 - -> the-third-librarian-never-left-you 3.41s on-27 - -> what-to-take-with-you 6.52s on-28 - -> where-the-file-came-from 5.82s on-29 - -> top 30.24s on-30 - -> the-first-promise-your-visit-stays 3.91s on-31 - -> the-thorough-walk 3.17s on-32 - -> receipts 3.38s on-33 - -> top 12.45s on-34 - -> top 7.50s on-35 - -> vendors 4.95s on-36 - -> top 5.25s on-37 - -> the-instrument 3.92s on-38 - -> provenance-4cae43 5.01s on-39 - -> provenance-063a00 4.22s on-40 - -> the-blind-sheet-unsealed 7.94s on-41 - -> the-blind-sitting 4.81s on-42 ['no template tail — no closing sentence of the shape "the article does not specify what was left out, moved or not measured" where the article carries no such statement', "an operator, never the definite one — the hub's voice law (an operator 2026-09-03, re-stated 2026-09-04)", "the prediction's own subject kept — the page predicts that a few COMPANIES will sell intelligence as a utility; the commodity sentence is about SOLD INTELLIGENCE and may not be re-subjected onto the prediction"] -> what-an-operator-saw 3.82s {"n": 42, "n_labelled": 1, "correct": 0, "errors": 0, "brier": 1.0, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": n ANNOUNCE box benchbox · card GPU-46890836 · arm base-generate · task c · start 2026-09-21T17:56:12Z c-darwin-01-dessert-01 darwin -> darwin 1.57s c-darwin-01-dessert-01 darwin -> darwin 1.04s c-darwin-01-dessert-01 darwin -> ibn_sina 1.05s c-darwin-01-dessert-02 darwin -> darwin 1.05s c-darwin-01-dessert-03 darwin -> darwin 1.12s c-darwin-01-dessert-03 darwin -> darwin 1.08s c-darwin-01-main-01 darwin -> darwin 1.08s c-darwin-01-main-02 darwin -> ibn_sina 1.10s c-darwin-01-main-02 darwin -> darwin 1.05s c-darwin-01-main-02 darwin -> darwin 1.09s c-darwin-01-main-02 darwin -> darwin 1.00s c-darwin-01-main-03 darwin -> darwin 1.08s c-darwin-01-starter-01 darwin -> darwin 1.04s c-darwin-01-starter-01 darwin -> darwin 1.01s c-darwin-01-starter-01 darwin -> darwin 1.04s c-darwin-01-starter-02 darwin -> darwin 1.11s c-darwin-01-starter-03 darwin -> darwin 1.09s c-darwin-01-starter-03 darwin -> darwin 1.11s c-einstein-01-dessert-01 einstein -> sagan 1.03s c-einstein-01-dessert-01 einstein -> sagan 1.08s c-einstein-01-dessert-02 einstein -> einstein 1.06s c-einstein-01-dessert-02 einstein -> einstein 0.98s c-einstein-01-dessert-03 einstein -> einstein 1.10s c-einstein-01-main-01 einstein -> socrates 1.08s c-einstein-01-main-01 einstein -> einstein 1.08s c-einstein-01-main-02 einstein -> einstein 1.11s c-einstein-01-main-03 einstein -> sagan 1.05s c-einstein-01-main-03 einstein -> einstein 1.06s c-einstein-01-main-03 einstein -> einstein 1.05s c-einstein-01-main-03 einstein -> einstein 1.07s c-einstein-01-starter-01 einstein -> einstein 1.05s c-einstein-01-starter-02 einstein -> einstein 1.11s c-einstein-01-starter-02 einstein -> einstein 1.07s c-einstein-01-starter-03 einstein -> einstein 1.02s c-einstein-01-starter-03 einstein -> einstein 1.02s c-einstein-01-starter-03 einstein -> einstein 1.06s c-hypatia-01-dessert-01 hypatia -> hypatia 1.01s c-hypatia-01-dessert-02 hypatia -> hypatia 1.11s c-hypatia-01-dessert-02 hypatia -> sagan 1.01s c-hypatia-01-dessert-03 hypatia -> ibn_sina 1.04s c-hypatia-01-main-01 hypatia -> hypatia 1.06s c-hypatia-01-main-01 hypatia -> einstein 1.08s c-hypatia-01-main-01 hypatia -> hypatia 1.07s c-hypatia-01-main-01 hypatia -> hypatia 1.05s c-hypatia-01-main-01 hypatia -> sagan 0.99s c-hypatia-01-main-02 hypatia -> einstein 0.98s c-hypatia-01-main-03 hypatia -> einstein 1.12s c-hypatia-01-main-03 hypatia -> hypatia 1.07s c-hypatia-01-main-03 hypatia -> hypatia 1.02s c-hypatia-01-starter-01 hypatia -> hypatia 1.08s c-hypatia-01-starter-01 hypatia -> socrates 1.05s c-hypatia-01-starter-02 hypatia -> hypatia 1.02s c-hypatia-01-starter-02 hypatia -> hypatia 1.07s c-hypatia-01-starter-02 hypatia -> hypatia 0.99s c-ibn_sina-01-dessert-01 ibn_sina -> sagan 1.13s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 1.01s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 1.05s c-ibn_sina-01-dessert-02 ibn_sina -> ibn_sina 0.97s c-ibn_sina-01-dessert-03 ibn_sina -> ibn_sina 1.06s c-ibn_sina-01-dessert-03 ibn_sina -> ibn_sina 1.08s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.08s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.04s c-ibn_sina-01-main-02 ibn_sina -> ibn_sina 1.07s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.11s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.06s c-ibn_sina-01-main-03 ibn_sina -> ibn_sina 1.04s c-ibn_sina-01-starter-01 ibn_sina -> hypatia 1.11s c-ibn_sina-01-starter-02 ibn_sina -> socrates 0.96s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.02s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.08s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.08s c-ibn_sina-01-starter-03 ibn_sina -> ibn_sina 1.15s c-sagan-01-dessert-01 sagan -> sagan 1.08s c-sagan-01-dessert-01 sagan -> sagan 1.04s c-sagan-01-dessert-01 sagan -> sagan 1.02s c-sagan-01-dessert-03 sagan -> sagan 1.02s c-sagan-01-dessert-03 sagan -> sagan 1.12s c-sagan-01-dessert-03 sagan -> sagan 1.05s c-sagan-01-dessert-03 sagan -> sagan 1.07s c-sagan-01-dessert-03 sagan -> sagan 1.08s c-sagan-01-main-01 sagan -> sagan 1.00s c-sagan-01-main-02 sagan -> sagan 1.07s c-sagan-01-main-02 sagan -> sagan 1.05s c-sagan-01-main-03 sagan -> sagan 1.08s c-sagan-01-main-03 sagan -> sagan 1.09s c-sagan-01-starter-01 sagan -> sagan 1.06s c-sagan-01-starter-02 sagan -> sagan 1.07s c-sagan-01-starter-02 sagan -> sagan 0.96s c-sagan-01-starter-03 sagan -> darwin 1.09s c-sagan-01-starter-03 sagan -> einstein 1.03s c-socrates-01-dessert-01 socrates -> socrates 1.11s c-socrates-01-dessert-01 socrates -> socrates 0.96s c-socrates-01-dessert-01 socrates -> socrates 1.05s c-socrates-01-dessert-02 socrates -> sagan 1.07s c-socrates-01-dessert-02 socrates -> socrates 0.95s c-socrates-01-dessert-02 socrates -> socrates 1.06s c-socrates-01-main-01 socrates -> socrates 1.01s c-socrates-01-main-02 socrates -> socrates 1.10s c-socrates-01-main-02 socrates -> socrates 1.04s c-socrates-01-main-02 socrates -> socrates 1.08s c-socrates-01-main-02 socrates -> socrates 1.05s c-socrates-01-main-03 socrates -> socrates 0.96s c-socrates-01-main-03 socrates -> socrates 1.06s c-socrates-01-starter-01 socrates -> socrates 1.03s c-socrates-01-starter-02 socrates -> socrates 1.01s c-socrates-01-starter-03 socrates -> socrates 1.08s c-socrates-01-starter-03 socrates -> socrates 1.09s c-socrates-01-starter-03 socrates -> socrates 1.04s {"n": 108, "n_labelled": 108, "correct": 89, "errors": 0, "brier": 0.35956790123456794, "reliability": [{"lo": 0.0, "hi": 0.1, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.1, "hi": 0.2, "n": 1, "sum_p": 0.16666666666666666, "correct": 1, "mean_p": 0.16666666666666666, "accuracy": 1.0, "gap": -0.8333333333333334}, {"lo": 0.2, "hi": 0.3, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.3, "hi": 0.4, "n": 0, "sum_p": 0.0, "correct": 0, "mean_p": null, "accuracy": null, "gap": null}, {"lo": 0.4, "hi": 0.5,