{"id": "on-01", "slug": "three-new-voices-at-the-narrators-chair", "label": "yes", "kind": "on-page", "question": "what did the round cost in total, and how were the three new voices judged?", "n_chunks": 290, "score": -0.6114794611930847, "best_heading": "What changed since the open call {#what-changed-since-the-open-call}", "seconds": 3.9139014000538737}
{"id": "on-02", "slug": "two-new-frontier-models-at-the-rules-desk", "label": "yes", "kind": "on-page", "question": "how many cases did the blind head-to-head use, and did it separate the two models?", "n_chunks": 120, "score": 6.202847480773926, "best_heading": "The short version {#top}", "seconds": 1.4054326789919287}
{"id": "on-03", "slug": "six-worlds-painted-large", "label": "yes", "kind": "on-page", "question": "how many worlds were painted, and at what size and step count?", "n_chunks": 17, "score": 0.5422636270523071, "best_heading": "The short version {#top}", "seconds": 0.31982539198361337}
{"id": "on-04", "slug": "the-beat-lab-asked-and-answered", "label": "yes", "kind": "on-page", "question": "how many answers does the page carry, and does any model run to make it?", "n_chunks": 36, "score": 5.1186418533325195, "best_heading": "Is this AI? {#is-this-ai}", "seconds": 0.339427919825539}
{"id": "on-05", "slug": "ten-minutes-with-living-artists", "label": "yes", "kind": "on-page", "question": "how many tracks and releases went into the adapter, and what did the one listener prefer?", "n_chunks": 56, "score": 4.666945457458496, "best_heading": "What to take with you {#what-to-take-with-you}", "seconds": 0.5867862130980939}
{"id": "on-06", "slug": "listen-for-yourself", "label": "yes", "kind": "on-page", "question": "how many renders are on the page, and what was done to their loudness?", "n_chunks": 45, "score": 1.846085548400879, "best_heading": "How to listen {#how-to-listen}", "seconds": 0.7521048970520496}
{"id": "on-07", "slug": "half-an-hour-with-dead-composers", "label": "yes", "kind": "on-page", "question": "how long did the Sousa and Bach adapters take to train?", "n_chunks": 27, "score": 3.735194444656372, "best_heading": "Previously, in five minutes {#previously-in-five-minutes}", "seconds": 0.18917703395709395}
{"id": "on-08", "slug": "chopin-in-five-minutes", "label": "yes", "kind": "on-page", "question": "how many recordings were used, and how long did the training take by the trainer's clock?", "n_chunks": 20, "score": 2.3094756603240967, "best_heading": "Five minutes of training, honestly accounted {#five-minutes-of-training-honestly-accounted}", "seconds": 0.13554142299108207}
{"id": "on-09", "slug": "the-same-sixteen", "label": "yes", "kind": "on-page", "question": "what did glm-5.3-flash score on the tool bench, and what did the local models score?", "n_chunks": 33, "score": 6.064107894897461, "best_heading": "The short version {#top}", "seconds": 0.2925241789780557}
{"id": "on-10", "slug": "two-hours-on-battery", "label": "yes", "kind": "on-page", "question": "what speed did the mini pc hold on battery, and for how long?", "n_chunks": 49, "score": 6.265389442443848, "best_heading": "Two hours at 12 tok/s \u2014 the test that named the page {#two-hours-at-12-tok-s-the-test-that-named-the-page}", "seconds": 0.6551595388446003}
{"id": "on-11", "slug": "the-instrument-travels", "label": "yes", "kind": "on-page", "question": "when the frozen exam moved to the smaller card, where did the ranked model land?", "n_chunks": 38, "score": 0.778886616230011, "best_heading": "What to take with you {#what-to-take-with-you}", "seconds": 0.46468359394930303}
{"id": "on-12", "slug": "nine-worlds-one-dog", "label": "yes", "kind": "on-page", "question": "what stays the same across the nine paintings?", "n_chunks": 8, "score": -1.7797985076904297, "best_heading": "The short version {#top}", "seconds": 0.0802895228844136}
{"id": "on-13", "slug": "what-150-watts-buys", "label": "yes", "kind": "on-page", "question": "what did the 450-watt cap cost the dense model, and what did it buy back?", "n_chunks": 34, "score": 5.374271392822266, "best_heading": "The answer, in one paragraph {#the-answer-in-one-paragraph}", "seconds": 0.3803338350262493}
{"id": "on-14", "slug": "how-the-beat-lab-works", "label": "yes", "kind": "on-page", "question": "where is the sound made, and how long does a typical wish take?", "n_chunks": 169, "score": 1.794629693031311, "best_heading": "What to take with you {#what-to-take-with-you}", "seconds": 1.2187281139194965}
{"id": "on-15", "slug": "the-typist-and-the-developer", "label": "yes", "kind": "on-page", "question": "how long did the developer take to develop one picture?", "n_chunks": 28, "score": 2.777064323425293, "best_heading": "One card, both jobs, at once {#one-card-both-jobs-at-once}", "seconds": 0.3939285599626601}
{"id": "on-16", "slug": "a-rig-your-friend-already-owns", "label": "yes", "kind": "on-page", "question": "how much faster was the workstation card than the 3090 on the game's own art?", "n_chunks": 22, "score": 4.378505706787109, "best_heading": "What this means for the game {#what-this-means-for-the-game}", "seconds": 0.2112209121696651}
{"id": "on-17", "slug": "how-a-vision-model-sees", "label": "yes", "kind": "on-page", "question": "how many tokens does an 896 px picture become, and how long does its prefill take?", "n_chunks": 52, "score": 4.909300804138184, "best_heading": "Finding one: a picture is 199 tokens at 896 px, and the runtime says so \u2014 on one seat {#finding-one-a-picture-is-199-tokens-at-896-px-and-the}", "seconds": 0.4230886909645051}
{"id": "on-18", "slug": "fifty-seven-milliseconds", "label": "yes", "kind": "on-page", "question": "which painter is the fastest cell, and at what time?", "n_chunks": 23, "score": 3.6617045402526855, "best_heading": "Finding four: the argument with our own settings \u2014 and what it changed {#finding-four-the-argument-with-our-own-settings-and-what-it}", "seconds": 0.30383400386199355}
{"id": "on-19", "slug": "the-dog-the-dice-and-the-painter", "label": "yes", "kind": "on-page", "question": "how big is the file the whole world is written in?", "n_chunks": 30, "score": 1.2891902923583984, "best_heading": "The world is a file {#the-world-is-a-file}", "seconds": 0.30735595989972353}
{"id": "on-20", "slug": "where-new-knowledge-comes-from", "label": "yes", "kind": "on-page", "question": "what are the two engines the page names for where new knowledge comes from?", "n_chunks": 32, "score": -1.0432994365692139, "best_heading": "Two kinds of there-all-along {#two-kinds-of-there-all-along}", "seconds": 0.20672890916466713}
{"id": "on-21", "slug": "compression-is-the-objective", "label": "yes", "kind": "on-page", "question": "what does the page say a language model is, and what prize does it point to?", "n_chunks": 29, "score": 3.0551071166992188, "best_heading": "How to check our work \u2014 and see it live {#how-to-check-our-work-and-see-it-live}", "seconds": 0.2637668380048126}
{"id": "on-22", "slug": "the-free-speed-wasnt-free", "label": "yes", "kind": "on-page", "question": "what was the production median before the change, and what caused the speed-up?", "n_chunks": 35, "score": -3.8681626319885254, "best_heading": "The short version {#top}", "seconds": 0.35985498991794884}
{"id": "on-23", "slug": "the-compressed-photograph", "label": "yes", "kind": "on-page", "question": "how much slower was the uncompressed original than the compressed build?", "n_chunks": 42, "score": 5.689628601074219, "best_heading": "What to take with you {#what-to-take-with-you}", "seconds": 0.40669191209599376}
{"id": "on-24", "slug": "three-at-the-table", "label": "yes", "kind": "on-page", "question": "what is the true active share for the live seat, and what does the page call the two numbers in a model name?", "n_chunks": 37, "score": 3.1406476497650146, "best_heading": "The payroll and the meeting {#the-payroll-and-the-meeting}", "seconds": 0.3562020240351558}
{"id": "on-25", "slug": "reading-is-fast", "label": "yes", "kind": "on-page", "question": "what was the median reading speed, and why is reading faster than writing?", "n_chunks": 29, "score": 3.714657783508301, "best_heading": "The mystery, stated honestly {#the-mystery-stated-honestly}", "seconds": 0.3222333509474993}
{"id": "on-26", "slug": "three-librarians", "label": "yes", "kind": "on-page", "question": "how many dimensions does the embedder's map have, and what does the third librarian do?", "n_chunks": 31, "score": 0.039186351001262665, "best_heading": "The second librarian searches by meaning {#the-second-librarian-searches-by-meaning}", "seconds": 0.2425428070127964}
{"id": "on-27", "slug": "the-new-kid", "label": "yes", "kind": "on-page", "question": "which seat did the new model pass, and which floor did it miss?", "n_chunks": 67, "score": 2.980286121368408, "best_heading": "The judge seat \u2014 and the floor it missed {#the-judge-seat-and-the-floor-it-missed}", "seconds": 0.6910471338778734}
{"id": "on-28", "slug": "llms-txt", "label": "yes", "kind": "on-page", "question": "did any ai crawler fetch the file in the first thirty days, and how big is the largest measured llms-full file?", "n_chunks": 58, "score": 1.8106272220611572, "best_heading": "Measured again, after publication {#measured-again-after-publication}", "seconds": 0.7145207030698657}
{"id": "on-29", "slug": "the-open-call", "label": "yes", "kind": "on-page", "question": "how many models were handed the frozen moment, and what did the round cost?", "n_chunks": 290, "score": 3.2757363319396973, "best_heading": "The short version {#top}", "seconds": 2.696088836994022}
{"id": "on-30", "slug": "where-your-question-goes", "label": "yes", "kind": "on-page", "question": "how long does every notebook on the server keep its entries?", "n_chunks": 27, "score": 6.202828407287598, "best_heading": "The first promise: your visit stays between us {#the-first-promise-your-visit-stays-between-us}", "seconds": 0.22016151691786945}
{"id": "on-31", "slug": "how-we-work", "label": "yes", "kind": "on-page", "question": "what does the page say about the humans and agents who make each exhibit?", "n_chunks": 27, "score": 4.606722831726074, "best_heading": "The short version {#top}", "seconds": 0.2285258339252323}
{"id": "on-32", "slug": "the-move", "label": "yes", "kind": "on-page", "question": "what did the median search time go from and to after the move?", "n_chunks": 30, "score": 0.719709038734436, "best_heading": "Instrument two \u00b7 the receipts {#receipts}", "seconds": 0.41453083301894367}
{"id": "on-33", "slug": "chair-trials", "label": "yes", "kind": "on-page", "question": "how many models and arms sat the five exams, and did the box stop serving?", "n_chunks": 116, "score": 6.620803356170654, "best_heading": "The short version {#top}", "seconds": 1.6483579569030553}
{"id": "on-34", "slug": "cove-voice-head-to-head", "label": "yes", "kind": "on-page", "question": "what happened to the August class's best voice at the cove's gate?", "n_chunks": 69, "score": 3.649540901184082, "best_heading": "The short version {#top}", "seconds": 0.8967985219787806}
{"id": "on-35", "slug": "outside-judges", "label": "yes", "kind": "on-page", "question": "how many outside labs re-judged, and what was the widest cross-vendor gap?", "n_chunks": 45, "score": 4.444449424743652, "best_heading": "What to take with you {#what-to-take-with-you}", "seconds": 0.638277105987072}
{"id": "on-36", "slug": "august-arrivals", "label": "yes", "kind": "on-page", "question": "how many models landed that week, and what share broke the response contract?", "n_chunks": 48, "score": 0.8004065752029419, "best_heading": "What to take with you {#what-to-take-with-you}", "seconds": 0.7003170819953084}
{"id": "on-37", "slug": "seat-trials", "label": "yes", "kind": "on-page", "question": "what are the kill-recall and preservation floors a model must clear?", "n_chunks": 30, "score": 4.524374485015869, "best_heading": "What to take with you {#what-to-take-with-you}", "seconds": 0.40392806800082326}
{"id": "on-38", "slug": "voice-trials", "label": "yes", "kind": "on-page", "question": "how many model tags were tried, and what does the anchor model read?", "n_chunks": 43, "score": 4.507524013519287, "best_heading": "What to take with you {#what-to-take-with-you}", "seconds": 1.2570502532180399}
{"id": "on-39", "slug": "licences", "label": "yes", "kind": "on-page", "question": "what does the page say it does with each model's licence?", "n_chunks": 35, "score": 3.8678011894226074, "best_heading": "How to read this page {#how-to-read-this-page}", "seconds": 0.5051935070659965}
{"id": "on-40", "slug": "hear-it-for-yourself", "label": "yes", "kind": "on-page", "question": "how many clips does the page carry, and how many presentations made up the sitting?", "n_chunks": 57, "score": 2.5990633964538574, "best_heading": "The blind sheet, unsealed {#the-blind-sheet-unsealed}", "seconds": 0.7289199589286}
{"id": "on-41", "slug": "the-ceiling-is-not-the-corpus", "label": "yes", "kind": "on-page", "question": "how did the adapter score against the base model in the listening, and on how many house prompts did it win?", "n_chunks": 50, "score": 2.8058533668518066, "best_heading": "What the adapter actually did {#what-the-adapter-actually-did}", "seconds": 0.4743491169065237}
{"id": "on-42", "slug": "the-indifferent-prediction", "label": "yes", "kind": "on-page", "question": "what's the prediction?", "n_chunks": 39, "score": 1.458088755607605, "best_heading": "The indifferent prediction {#the-indifferent-prediction}", "seconds": 0.22659760201349854}
{"id": "off-01", "slug": "what-150-watts-buys", "label": "no", "kind": "off-page", "question": "what does a large cloud provider charge for the same work?", "n_chunks": 34, "score": -10.350339889526367, "best_heading": "(before the first heading)", "seconds": 0.32509227795526385}
{"id": "off-02", "slug": "how-we-work", "label": "no", "kind": "off-page", "question": "what is the postal address of the workshop that ran this?", "n_chunks": 27, "score": -3.6086368560791016, "best_heading": "How to check our work \u2014 and see it live {#how-to-check-our-work-and-see-it-live}", "seconds": 0.22044044896028936}
{"id": "off-03", "slug": "llms-txt", "label": "no", "kind": "off-page", "question": "how many people subscribe to this site's feed?", "n_chunks": 58, "score": -7.991389274597168, "best_heading": "The per-arm table \u2014 no pooled figure prints without it {#the-per-arm-table-no-pooled-figure-prints-without-it}", "seconds": 0.6079953168518841}
{"id": "off-04", "slug": "three-new-voices-at-the-narrators-chair", "label": "no", "kind": "off-page", "question": "what will the next article on this site be about?", "n_chunks": 290, "score": -9.983488082885742, "best_heading": "What this page does not say {#what-this-page-does-not-say}", "seconds": 3.5772057259455323}
{"id": "off-05", "slug": "two-hours-on-battery", "label": "no", "kind": "off-page", "question": "what did this cost in electricity at my own utility's rate?", "n_chunks": 49, "score": -10.013277053833008, "best_heading": "(before the first heading)", "seconds": 0.7658131229691207}
{"id": "off-06", "slug": "reading-is-fast", "label": "no", "kind": "off-page", "question": "who won the football world cup in 2022?", "n_chunks": 29, "score": -10.897608757019043, "best_heading": "The number that argued back {#the-number-that-argued-back}", "seconds": 0.45972436387091875}
{"id": "off-07", "slug": "a-rig-your-friend-already-owns", "label": "no", "kind": "off-page", "question": "what does one of these cards cost in my country today?", "n_chunks": 22, "score": -6.8804473876953125, "best_heading": "Three findings we didn't order {#three-findings-we-didnt-order}", "seconds": 0.20661279186606407}
{"id": "off-08", "slug": "the-move", "label": "no", "kind": "off-page", "question": "can you write a python script that reproduces this for me?", "n_chunks": 30, "score": -9.709053993225098, "best_heading": "How to check our work \u2014 and see it live {#how-to-check-our-work-and-see-it-live}", "seconds": 0.41347977705299854}
{"id": "off-09", "slug": "fifty-seven-milliseconds", "label": "no", "kind": "off-page", "question": "what was the weather like on the day this was measured?", "n_chunks": 23, "score": -10.1728515625, "best_heading": "Finding one: a step is a fixed-price good {#finding-one-a-step-is-a-fixed-price-good}", "seconds": 0.43032786599360406}
{"id": "off-10", "slug": "six-worlds-painted-large", "label": "no", "kind": "off-page", "question": "how many hours a week do the people who ran this work?", "n_chunks": 17, "score": -9.818023681640625, "best_heading": "(before the first heading)", "seconds": 0.12411465891636908}
{"id": "off-11", "slug": "outside-judges", "label": "no", "kind": "off-page", "question": "what is this workshop's hourly rate for consulting work?", "n_chunks": 45, "score": -7.261738300323486, "best_heading": "What it cost {#bill}", "seconds": 0.5929714241065085}
{"id": "off-12", "slug": "the-typist-and-the-developer", "label": "no", "kind": "off-page", "question": "what is this workshop's annual revenue?", "n_chunks": 28, "score": -11.312265396118164, "best_heading": "The rest of the seminar {#the-rest-of-the-seminar}", "seconds": 0.31805749516934156}
{"id": "off-13", "slug": "compression-is-the-objective", "label": "no", "kind": "off-page", "question": "how would these numbers change on windows instead?", "n_chunks": 29, "score": -9.181620597839355, "best_heading": "What we can show you {#what-we-can-show-you}", "seconds": 0.31178234005346894}
{"id": "off-14", "slug": "the-open-call", "label": "no", "kind": "off-page", "question": "what did other reviewers say about this piece after it was published?", "n_chunks": 290, "score": -2.265803098678589, "best_heading": "A late contestant, and it ties {#a-late-contestant-and-it-ties}", "seconds": 3.251832216978073}
{"id": "off-15", "slug": "the-new-kid", "label": "no", "kind": "off-page", "question": "how does this compare to a result published in 2028?", "n_chunks": 67, "score": 0.817664384841919, "best_heading": "The judge seat \u2014 and the floor it missed {#the-judge-seat-and-the-floor-it-missed}", "seconds": 0.8075895719230175}
{"id": "off-16", "slug": "where-your-question-goes", "label": "no", "kind": "off-page", "question": "which country is the workshop's business registered in?", "n_chunks": 27, "score": -9.08049201965332, "best_heading": "The rest of the seminar {#the-rest-of-the-seminar}", "seconds": 0.24172906088642776}
{"id": "off-17", "slug": "nine-worlds-one-dog", "label": "no", "kind": "off-page", "question": "how many other readers have asked this same question?", "n_chunks": 8, "score": -9.38121223449707, "best_heading": "For the technically curious {#for-the-technically-curious}", "seconds": 0.07752310507930815}
{"id": "off-18", "slug": "how-the-beat-lab-works", "label": "no", "kind": "off-page", "question": "what phone number reaches support?", "n_chunks": 169, "score": -10.694653511047363, "best_heading": "What to take with you {#what-to-take-with-you}", "seconds": 1.2902673750650138}
{"id": "off-19", "slug": "three-at-the-table", "label": "no", "kind": "off-page", "question": "what would this measurement look like at twice the batch size?", "n_chunks": 37, "score": -7.231453895568848, "best_heading": "What sparsity does not buy {#what-sparsity-does-not-buy}", "seconds": 0.36286658188328147}
{"id": "off-20", "slug": "the-same-sixteen", "label": "no", "kind": "off-page", "question": "which of these models is the best one to use for my own project?", "n_chunks": 33, "score": -4.3440093994140625, "best_heading": "A different kind of row {#a-different-kind-of-row}", "seconds": 0.41378539404831827}
{"id": "off-21", "slug": "seat-trials", "label": "no", "kind": "off-page", "question": "how many models will sit this exam next year?", "n_chunks": 30, "score": 0.12131539732217789, "best_heading": "The short version {#top}", "seconds": 0.43273022794164717}