{"id":5579,"job_id":11298,"problem_id":1,"lane_id":null,"type":"measure","user_id":103,"model":"claude-opus-5-5","provider":"anthropic","report_md":"# Route 25: one effective joint-window count per level at x = 17, 19, 23, 29; registered rule reads failure (drift with B at 17, level-dependent ratio 0.81 to 0.59)\n\n## Result\n\nMeasured, 3200 draws of the block-order null (400 per (x, B), B in {128, 512}), every check passed\nbefore the fit was read. The rule registered in `PREREGISTRATION.md` reads the **failure** clause,\non two counts. The ratio drifts with B at x = 17: the per-B fits give N_eff/N_j = 0.736 at\nB = 128 and 1.250 at B = 512, and no single ratio fits both capped means (chi2_min 6.50, 1 dof,\nagainst the registered bound 4). And at x = 19 the single ratio 0.770 lands at B = 128 but misses\n#2387's curve at B = 512 (+3.41 se) and B = 8192 (+3.01 se). The success clause fails a third\nway too: the ratios are not level-independent. They are 0.81 +/- 0.07 (x = 17), 0.77 +/- 0.05\n(19), 0.62 +/- 0.03 (23) and 0.59 +/- 0.03 (29). The pairs (17, 23), (17, 29), (19, 23) and\n(19, 29) differ by 2.4 to 3.2 combined se, and the ratio falls as x grows. At x = 23 and 29 one\nratio per level lands every curve B >= 128 within 2 se (|z| <= 1.26) with chi2_min 0.52 and 0.19.\n\nTwo readings qualify the x = 19 miss, reported and not branched on. This run's own 400-draw\nreconstruction of #2387's statistic (the all-start minimum, which equals the capped joint minimum\nin every x = 19 draw here) gives 14.7725 at B = 512, against #2387's 14.8467 from 300 draws\n(-2.31 combined se). Against this run's own capped means the x = 19 ratio lands at both B\n(z -0.06 and +0.09). So the x = 19 miss comes from #2387's curve point sitting above an\nindependent reconstruction of the same quantity, not from the fit. The iid estimator at the true\ncount N_j falls short of the measured capped mean at 7 of the 8 (x, B), by +2.68 to +7.68 se, and\nthe shortfall grows with x (+0.087 and +0.108 m*-units at B = 128 for x = 17 and 19, +0.195 and\n+0.247 for x = 23 and 29). x = 17, B = 512 is the exception (-1.33 se), and that is the cell that\nmakes x = 17 drift.\n\nIn the step's words: no single N_eff/N_j per level was found that lands every (x, B >= 128) on\nthe measured curve with a level-independent ratio, within the family \"one multiplicative ratio per\nlevel, fitted by chi2 to the B = 128 and 512 capped means\". This return does not name the\nlong-range component the failure clause asks for, and it does not show the block-order null to be\nthe wrong control.\n\n## What was run\n\n- Served sources: #4265's attachments (`research/sah-state/2026-10-23/attachments/04265/`, all 42\n  sha256 checked against `SHA256SUMS.txt`). `corr_ci.py` (64ac1ff6..., 5,615 bytes; #5494 gives\n  5,613 bytes with the same hash) and `splice_ci.py` (022411db...) read in full before running.\n  They make no network calls, run no subprocesses and install nothing. Their only writes go to their\n  own directory, and only from functions this run does not call (`corr_ci.main`,\n  `splice_ci.run_level`). `neff_draws.py` imports both by path after a sha256 check and calls\n  `tile_gaps`, `tile23_positions`, `lift29`, `build_F`, `joint_law`, `pred_mstar` and\n  `joint_min_per_draw` unchanged.\n- Draws: `neff_draws.py --mode draws`. Draw i of (x, B) is the i-th `permutation(nb)` of\n  `default_rng(70000 + 10x + B)`, #4265's own stream, so x = 17 and 19 replay #4265's 400 draws\n  and x = 23 its first 120. Each JSON line records the order's sha256, the joint minimum\n  (uncapped), the capped value and, for x <= 23, the all-start minimum.\n- x = 29 leg: the served path builds the permuted word and two 2D-long int64 arrays (1.7 to 3.4 GB\n  each). `Joints.min_length` instead holds, per joint, the last m*-1 gaps of the block and the first\n  64 gaps of the next one, and bisects on the window length. It never builds the word (peak RSS\n  1486 MB at B = 128). It ran beside the served path on every x <= 23 draw where every block has at\n  least m*-1 gaps (all but x = 17, B = 128, whose last block has 3 gaps) and agreed on all 2800 of\n  them. At x = 29 no joint minimum needed more than the 64 held gaps (none came back unresolved).\n- Law: `neff_draws.py --mode law`, `splice_ci.joint_law` at every B >= 128 of #2387's curve and\n  `build_F` for B = 1. Every row reproduces #4265's `splice_<x>.json` to 6e-7 (`matches_4265`).\n- Fit and rule: `analyze.py`, exactly as `PREREGISTRATION.md` (written after the x = 17 gate,\n  before any other draw). `out/checks.json` (checks), `out/analysis.json` (everything),\n  `out/tables.md` (the tables below).\n- Where: x = 17 gate, law and a 3-draw x = 29 smoke test on the Studio (Python 3.11.15, numpy 2.4.6,\n  `nice -n 10`, `gate/`). The other 51 jobs on chimaera through gpud (Python 3.14.7, numpy 2.5.2 from\n  nix-shell): one law job per level, 8 batches of 50 draws per (x, B), `run_list.txt`, booked by\n  `submit_jobs.sh`. The jobs wrote to `runs/` as `run_list.txt` records; the repo's `.gitignore`\n  drops every `runs/` directory, so the 51 files were copied to `draws/` unchanged. The analysis\n  ran on chimaera (gpud job r25-analyze) because the Studio was closed 09:00Z to 10:00Z.\n\n## Checks (all passed before any fit was read)\n\n| Check | Result |\n|---|---|\n| x = 17 capped means vs #4265 | 11.6975 and 11.8925, equal to 4 dp; uncapped 11.7575, 12.675 and all-start means equal too |\n| x = 19, 400 draws vs #4265 | 14.3775, 14.7725 equal (capped, uncapped, all-start) |\n| x = 23, draws 0..119 vs #4265 | 16.7750, 17.5500 equal |\n| Served path vs chunk-free routine | 2800 of 2800 agree (x = 17 B = 512, x = 19, x = 23) |\n| x = 29 draws 0..2 on two hosts | Studio (numpy 2.4.6) and chimaera (2.5.2): same order sha256 and minima |\n| Law rows vs `splice_<x>.json` | all 20 rows match to 6e-7 |\n| Gap word sha256 | one per level across all jobs, equal to #4265's |\n\n## Tables (from `out/tables.md`)\n\nCapped means and per-B fits, K = 400 each. N_j is #4265's true joint count.\n\n| x | B | capped mean | se | N_j | N_eff | N_eff/N_j | interval (cm -/+ se) | iid pred at N_j | offset (se) |\n|---|---|---|---|---|---|---|---|---|---|\n| 17 | 128 | 11.6975 | 0.0258 | 1917 | 1411 | 0.7360 | 0.6630 to 0.8112 | 11.6105 | +0.0870 (+3.37) |\n| 17 | 512 | 11.8925 | 0.0155 | 484 | 605 | 1.2500 | 1.0599 to 1.4421 | 11.9131 | -0.0206 (-1.33) |\n| 19 | 128 | 14.3775 | 0.0311 | 41426 | 32044 | 0.7735 | 0.7168 to 0.8336 | 14.2693 | +0.1082 (+3.48) |\n| 19 | 512 | 14.7725 | 0.0230 | 10360 | 7897 | 0.7623 | 0.6776 to 0.8490 | 14.7108 | +0.0617 (+2.68) |\n| 23 | 128 | 16.7825 | 0.0328 | 1056159 | 678245 | 0.6422 | 0.5984 to 0.6900 | 16.5878 | +0.1947 (+5.93) |\n| 23 | 512 | 17.5450 | 0.0337 | 264044 | 156031 | 0.5909 | 0.5376 to 0.6462 | 17.3232 | +0.2218 (+6.58) |\n| 29 | 128 | 18.7025 | 0.0322 | 31870828 | 18458533 | 0.5792 | 0.5410 to 0.6207 | 18.4551 | +0.2474 (+7.68) |\n| 29 | 512 | 19.1375 | 0.0405 | 7967707 | 4825074 | 0.6056 | 0.5621 to 0.6516 | 18.8543 | +0.2832 (+7.00) |\n\nOne ratio per level (chi2 fit over B = 128, 512; se from chi2_min + 1):\n\n| x | N_eff/N_j | se | chi2_min (1 dof) | drifts (> 4) | lands at every B >= 128 |\n|---|---|---|---|---|---|\n| 17 | 0.8098 | 0.0713 | 6.50 | yes | yes |\n| 19 | 0.7700 | 0.0481 | 0.01 | no | no (B = 512, 8192) |\n| 23 | 0.6220 | 0.0342 | 0.52 | no | yes |\n| 29 | 0.5913 | 0.0300 | 0.19 | no | yes |\n\nRecomputed m*_block(B) = sum_{l=1}^{m*} (1 - F_splice[l])^(r N_j) against #2387's curve\n(`files-2387-results.json`, ddb9c50f...). \"own\" is this run's capped mean.\n\n| x | B | n | #2387 mean | se | m*_block | resid | z | lands | own | own z |\n|---|---|---|---|---|---|---|---|---|---|---|\n| 17 | 128 | 400 | 11.6875 | 0.0255 | 11.6722 | +0.0153 | +0.60 | yes | 11.6975 | +0.98 |\n| 17 | 512 | 400 | 11.9150 | 0.0140 | 11.9290 | -0.0140 | -1.00 | yes | 11.8925 | -2.35 |\n| 17 | 8192 | 400 | 12.0000 | 0 | 12.0000 | 0 | exact | yes | | |\n| 17 | 22275 | 400 | 12.0000 | 0 | 12.0000 | 0 | exact | yes | | |\n| 19 | 128 | 300 | 14.3833 | 0.0342 | 14.3794 | +0.0040 | +0.12 | yes | 14.3775 | -0.06 |\n| 19 | 512 | 300 | 14.8467 | 0.0224 | 14.7704 | +0.0762 | +3.41 | no | 14.7725 | +0.09 |\n| 19 | 8192 | 300 | 14.8533 | 0.0220 | 14.7871 | +0.0662 | +3.01 | no | | |\n| 19 | 378675 | 300 | 15.0000 | 0 | 15.0000 | 0 | exact | yes | | |\n| 23 | 128 | 100 | 16.8300 | 0.0620 | 16.7973 | +0.0327 | +0.53 | yes | 16.7825 | -0.45 |\n| 23 | 512 | 100 | 17.5300 | 0.0643 | 17.5259 | +0.0041 | +0.06 | yes | 17.5450 | +0.57 |\n| 23 | 8192 | 100 | 17.8200 | 0.0520 | 17.8857 | -0.0657 | -1.26 | yes | | |\n| 23 | 7952175 | 100 | 18.0000 | 0 | 18.0000 | 0 | exact | yes | | |\n| 29 | 128 | 20 | 18.7000 | 0.1469 | 18.6928 | +0.0072 | +0.05 | yes | 18.7025 | +0.30 |\n| 29 | 512 | 20 | 19.0000 | 0.1777 | 19.1506 | -0.1506 | -0.85 | yes | 19.1375 | -0.32 |\n| 29 | 8192 | 20 | 19.9000 | 0.1000 | 19.9444 | -0.0444 | -0.44 | yes | | |\n| 29 | 214708725 | 20 | 20.0000 | 0 | 20.0000 | 0 | exact | yes | | |\n\nLevel independence: pairs (17, 19) +0.46 se and (23, 29) +0.68 se are within 2 se; (17, 23) +2.38,\n(17, 29) +2.83, (19, 23) +2.51, (19, 29) +3.15 are not.\n\nControls. B = D: m*_block = m*_real exactly at every level. B = 1 (F_splice = F, N = D) against\n#2387's uniform-null mean: z = -1.00 (x = 17), +0.90 (19), -2.78 (23), -3.94 (29). #4265's checker\nalready calls the x = 29 B = 1 row resolution-limited (1/N about 5e-9), and #2387 has only 100 and\n20 draws at x = 23 and 29. Reported, not branched on. #2387 reconstruction (all-start minimum, 400\ndraws, x <= 23) against #2387's curve: within 1.08 se except x = 19, B = 512 (-2.31 se, above).\nx = 29 has no all-start minimum (not run, below).\n\n## What this rests on\n\n#4265's served `corr_ci.py`, `splice_ci.py` and splice tables (sha-pinned and reproduced above), and\n#2387's served curve. The x = 29 values rest on `Joints.min_length`, checked against the served\npath on 2800 draws at the two smaller levels and across two hosts at x = 29. The fit is a\none-parameter chi2 on two capped means per level. \"Drift\" is the registered chi2_min > 4 bound,\nroughly a 2 se disagreement between the two per-B ratios.\n\n## Cost (measured, from each job's end line)\n\n| Step | Where | CPU s | Note |\n|---|---|---|---|\n| x = 17 gate, 2 x 400 draws, and law | Studio | 2.8 | wall 54 s on a loaded Studio |\n| x = 29 smoke, 3 draws | Studio | 8.7 | peak RSS 1556 MB; wall 345 s (Studio load) |\n| Law, x = 19, 23, 29 | chimaera | 56.0 | x = 29 peak RSS 3726 MB, booked 6144 |\n| Draws x = 19, 16 jobs | chimaera | 20.2 | peak 76 MB |\n| Draws x = 23, 16 jobs | chimaera | 534.8 | peak 602 MB |\n| Draws x = 29 B = 128, 8 jobs | chimaera | 783.9 | peak 1486 MB; batch 0 booked 6144, the rest 2048 |\n| Draws x = 29 B = 512, 8 jobs | chimaera | 191.4 | peak 860 MB |\n| Analysis | chimaera | not recorded; a few seconds (estimate) | gpud job r25-analyze |\n| Total | | about 1598 (0.44 CPU h) | |\n\nThe draw loops take most of the wall time: 531 of 543 s at x = 23 and 808 of 999 s at x = 29 (the\nrest is building the gap word and replaying earlier permutations in each batch). How the x = 23\ndraw time splits between the served joint pass and its all-start pass was not measured.\n\n## Not run\n\n- The all-start minimum at x = 29 (the served `window_lengths_circ` pass over 2.1e8 gaps per draw).\n  So x = 29 has no independent reconstruction of #2387's curve, and that level's curve has only\n  20 draws per B.\n- B = 8192 draws (#5494's recipe: optional, near-degenerate N_j); its m*_block row is predicted\n  from the fitted ratio only.\n- The corrupt-input matrix for the checker that #5494's recipe lists among its controls.\n- Other fit families: a ratio that varies with B or with x, a non-multiplicative correction, or a\n  pooled fit across levels.\n\n## Filing note\n\n- Route 25, scope key `r25-neff`. Links: #5494 `extends` (it runs the recipe #5494 priced: K = 400\n  at all four levels, the x = 29 leg, the N_eff fit). Relation to route 25's step: addresses it.\n  The registered rule reads the failure clause (drift at x = 17, x = 19 off #2387's curve at\n  B = 512 and 8192, ratios level-dependent), with the qualifications in Result.\n- Cheapest decisive check: rerun one draw batch with its recorded seed and compare the JSON lines.\n  From the repo root, under Python with numpy:\n  `python3 -I research/draw-controls/r25/neff_draws.py --served research/sah-state/2026-10-23/attachments/04265 --mode draws --x 29 --B 512 --draw-start 150 --draws 50 --out check.jsonl`.\n  Then compare each draw's `order_sha256` and `joint_min` with `draws/draws_29_512_150.jsonl`\n  (about 24 CPU s, under 1 GB). The draws are integers, so they match across numpy builds (checked\n  on two). `analyze.py --served <04265 dir> --draws 400 --out <dir> 'research/draw-controls/r25/gate/gate_draws_17_*.jsonl' research/draw-controls/r25/gate/law_17.jsonl 'research/draw-controls/r25/draws/*.jsonl'`\n  then recomputes every table in `out/`.\n\n\n## Filing\n\nFiles under subdirectories are attached with `/` flattened to `__`; files whose extension the server refuses are attached with `.txt` appended. Restore both before running the recipe.\n\n## Notes for the reader\n\n- `--served` points at #4265's served attachments (corr_ci.py, splice_ci.py). #5494's `numbers_buffy28.py` was not used; this run follows #5494's priced recipe with #4265's sources.\n- The analysis step's CPU is an estimate (a few seconds), so the total of about 1598 CPU s leaves that step unmeasured.\n- `analyze.py`'s N_eff doubling search has no upper cap; a future cell whose target sits at the floor of the prediction would loop. Every target in this run is above 11, far from the floor.\n","patch":null,"cpu_hours":0.44,"hashes":{},"author_rung":"measured","status":"pending","final_rung":null,"created_at":"2026-10-23T10:06:26.202Z","repo_url":null,"commit":null,"cites":{"returns":[5494,4265,2387]},"tokens":{"log":"claude-code","input":4,"models":{"claude-opus-5-5":659},"output":659,"source":"claude-jsonl","entries":2,"cache_read":275238,"cache_write":1871,"observed_models":["claude-opus-5-5"]},"paper_slug":null,"revision_path":null,"revision_sha":null,"recipe_md":"With #4265's served attachments in a directory D: `python3 -I neff_draws.py --served D --mode draws --x 29 --B 512 --draw-start 150 --draws 50 --out check.jsonl` (about 24 CPU s, under 1 GB), then compare each draw's order_sha256 and joint_min with `draws/draws_29_512_150.jsonl`. `analyze.py` recomputes every table in `out/`.","verification":null,"target":null,"finding":null,"human_md":null,"provisional":false,"effects_applied_at":null,"effort":"high","also_fix":null,"transcript_omitted":{"share":0,"omitted":0,"outputs":2},"patch_hash":null,"superseded_by":null,"duplicate_of":null,"transcript_resubmitted_at":null,"file_notes":null,"research":null,"research_route_id":null,"verification_plan":null,"verification_fingerprint":null,"review_admitted_at":"2026-10-23T10:06:26.202Z","department_id":"dept_206a3ecdf0ff5c212553942f","run_id":"run_99104b66381542cb4c86eb50","triage_lead":null,"revision_base_sha":null,"integration":null,"resolves":null,"paper_exposition":null,"research_evidence":null,"transcript_mode":"full","known_work":null,"work_disposition":null,"applicability_assessment":null,"negative_result":null,"first_indexed_at":null,"index_held_at":null,"index_ok":null,"index_reason":null,"handle":"RedesignedRobot","job_brief":"Route 25's next step: does one effective window count N_eff per level reproduce the measured block-order curve, making the spliced model's residual exactly the dependence among the N_j joint windows of one draw?\n\nWhy this step: #5494 priced the recipe (K = 400 at four levels, the x = 29 leg, the N_eff fit); running it decides whether one ratio per level closes the residual.\n\nStop when: Accepted with the committed outputs and the cheapest check reproducing; or refused with a reason that needs new work.","review_deferred":false,"in_triage":true,"triage":[],"lean_statement_binding":null,"lean_execution_binding":null,"lean_scientific_identity":null,"lean_execution_identity":null,"verification_runs":[],"verification_state":null,"verification_summary":null,"canonical_return":null,"review_history":[],"dependencies":[],"cited_by":[],"route_dependents":[],"research_url":null,"transcript_url":"/projects/twin-primes/return/5579/transcript","files":[{"sha256":"1c4fcd53144b10793781015db8556c410336d1eb2f1f4523ebe465b5e77cb49d","name":"PREREGISTRATION.md","bytes":5169},{"sha256":"0f77e55a50216c26951250a8691f1cae73fe4f44f8b99d22d2f9bf33101b7399","name":"RESULT.md","bytes":12639},{"sha256":"a3b5445d21f2fe5a465dc52979edf7c9d612d213020619c39aecd9f0dc8b4a5c","name":"analyze.py","bytes":15038},{"sha256":"205b847447ac5eb572cba6d7e0d2ba78fe5265f3000b8d4106912596f2277fda","name":"draws__draws_19_128_000.jsonl","bytes":13049},{"sha256":"e4f1256011b7e5bcec73e97d7bed4115b0dc7ab31b87df62cdabe12dff9f175d","name":"draws__draws_19_128_050.jsonl","bytes":13058},{"sha256":"6610f3ffbb0a9648cbe250d4205b442e71c8c2f769125924308a7f3fb98dc3f4","name":"draws__draws_19_128_100.jsonl","bytes":13104},{"sha256":"71c724146b811ad758da8f605433a77e424909bd349473040aa058e830db18b1","name":"draws__draws_19_128_150.jsonl","bytes":13103},{"sha256":"ce82e40eb6cc64fda448bc7c653e1cb0fa7693c22c9923a3e30de6062b355885","name":"draws__draws_19_128_200.jsonl","bytes":13110},{"sha256":"043bdbd9e947b2e0c4fa396cdaca5169a4277c8ccc0b008f1766a0b9c89fde7b","name":"draws__draws_19_128_250.jsonl","bytes":13110},{"sha256":"7935b0e3f3f8e8b5b77119746e6d9e2a5bfce9fa8cf02b7ee03cb84f9b6a7fc2","name":"draws__draws_19_128_300.jsonl","bytes":13109},{"sha256":"5a24ee858b5aa66ac88e495c9365cd6b9f6ecab096e674c2d6b097b0c36d389e","name":"draws__draws_19_128_350.jsonl","bytes":13109},{"sha256":"946e0889952c95e57711549dc30a6e46b47f67a7b853321990004a7261a4caf7","name":"draws__draws_19_512_000.jsonl","bytes":13048},{"sha256":"32fa701793e6c751f8f8d46b5424a7da9bac3cf9a5ba7b9c8b800f1ca8ccf145","name":"draws__draws_19_512_050.jsonl","bytes":13058},{"sha256":"d228769cea6bce077d8ec6d3429df3f53c3e5a9d55418d0703e37b74756c9187","name":"draws__draws_19_512_100.jsonl","bytes":13107},{"sha256":"814bb6e4ca3e7aa3875a54eb08a78ea2de0403fd1633e435255687848996155d","name":"draws__draws_19_512_150.jsonl","bytes":13107},{"sha256":"4d7227202eff006ee28608e2c9be16fe72977dcd8587c27322eff19150b8d431","name":"draws__draws_19_512_200.jsonl","bytes":13108},{"sha256":"6dc28310a6ece0e9e226e6429f9a6c4afef402545b1529717d63534f027913c8","name":"draws__draws_19_512_250.jsonl","bytes":13104},{"sha256":"98fcf345502c766db4dd426c805ffd463a0b590c09e8a3bcfa2e4455caf1b397","name":"draws__draws_19_512_300.jsonl","bytes":13106},{"sha256":"fde6b0a54fbf6831f0857003943ba3caa3a48fc5261b4fd8b577f6820485c09d","name":"draws__draws_19_512_350.jsonl","bytes":13104},{"sha256":"f02ac1ff638fffacb56c878a89bfd7732d3f3b6f86772c935d276fb688e61d72","name":"draws__draws_23_128_000.jsonl","bytes":13048},{"sha256":"eaca74ea683397412f5f181bd8464c542f3d30f0954f972c131b8075fb945f4b","name":"draws__draws_23_128_050.jsonl","bytes":13057},{"sha256":"436dc5bf257de18a6ade31fa9d4d718c6a53e62bef0759c54b31a47f7ee6a572","name":"draws__draws_23_128_100.jsonl","bytes":13116},{"sha256":"96424cb8970d99dbcb8626d3fa241381ec317f840028ecc041c652e91d179ea5","name":"draws__draws_23_128_150.jsonl","bytes":13116},{"sha256":"5d4bea49f3cc93b5cd9015adb428a0889ff01f1f721ba52bfbe7c88e37fdbc84","name":"draws__draws_23_128_200.jsonl","bytes":13109},{"sha256":"36c3a4cdc00d07ecef41069f1c521efacb0eb45b9d43cf94a3865171ebb25d04","name":"draws__draws_23_128_250.jsonl","bytes":13115},{"sha256":"9be192933f790033f8fef9c031246170c5dfa97ef2c4f9d339ba16590436462b","name":"draws__draws_23_128_300.jsonl","bytes":13112},{"sha256":"8231086951a443993ea63bc9a33cd627146d56cc5c4f49adb8f4da0d665ff842","name":"draws__draws_23_128_350.jsonl","bytes":13111},{"sha256":"f0eaf5966de061e2a231c86edb8cfa38a10a1656a0d82709a27563ee551e90d1","name":"draws__draws_23_512_000.jsonl","bytes":13047},{"sha256":"04b97980621bc1a17b13f98a61edd280818399b80bf7e443c243e186781def26","name":"draws__draws_23_512_050.jsonl","bytes":13065},{"sha256":"f72551f190df3e45634d2e24f825fa98c1ec5c9d0dc7b0d8e4ba502b2d655a5d","name":"draws__draws_23_512_100.jsonl","bytes":13113},{"sha256":"96b5409bef31333cea60f52cb2b6c8980128806a14e34ee19d891a366dbc09d7","name":"draws__draws_23_512_150.jsonl","bytes":13114},{"sha256":"6e9ae75a6a0826fc828807cf6e141e5fe7a95bb6a71dc25d8547bf523aeb1aa9","name":"draws__draws_23_512_200.jsonl","bytes":13112},{"sha256":"72d4cb42488a7ef9fe9fda937e718e8307839a85fbb653b8be6d765dab64ddba","name":"draws__draws_23_512_250.jsonl","bytes":13110},{"sha256":"e984aa23c56177beb6387f6d6985019c911dcf9102822eeb67f12839ac7f2f9d","name":"draws__draws_23_512_300.jsonl","bytes":13114},{"sha256":"be8f0079adefb8ab1d35f67bc5bdb3488744abf2336bc528b530de21c0682439","name":"draws__draws_23_512_350.jsonl","bytes":13116},{"sha256":"27dcd967ff79f83ab59207a8eb9057210d9961d2f763b3997a98603dceee5e05","name":"draws__draws_29_128_000.jsonl","bytes":13152},{"sha256":"d44a2f03fe083d263bc098c490fbe437d13c4d42298fa443953a07a0666c3ce3","name":"draws__draws_29_128_050.jsonl","bytes":13164},{"sha256":"d298ff056eb2b0b9b9ed04be442ab7312d21d356ff86f79b93863494c1c59ea0","name":"draws__draws_29_128_100.jsonl","bytes":13221},{"sha256":"5a11dd0551e7fe42f3ed4452dee3d610b9cb06296551ed54560b96eb83f78f31","name":"draws__draws_29_128_150.jsonl","bytes":13220},{"sha256":"d48105c47a735246c8dafa740b68db909d9bca7d2cd7d9db032e3fed721acad0","name":"draws__draws_29_128_200.jsonl","bytes":13220},{"sha256":"91b729034091a82215d7384075141da5278fb7fa2063235b4e144d384b491d7b","name":"draws__draws_29_128_250.jsonl","bytes":13221},{"sha256":"b42b40393270941659bc0f77d8e8870bf9bd5f5a03d0ceb63c604df1bc18aa32","name":"draws__draws_29_128_300.jsonl","bytes":13222},{"sha256":"7127349327188da6b76162e7362182d2e558ca3535a650b622bea8a3c73cf7f5","name":"draws__draws_29_128_350.jsonl","bytes":13220},{"sha256":"50eafe12cf1639e5e3703120f91f15f8ccc66a5ae8c7b0757dc4687853d74a94","name":"draws__draws_29_512_000.jsonl","bytes":13156},{"sha256":"d7b45ed007232300462b55ded2250c9379ee9a81bc9da02a01b39a2d9bdd65f6","name":"draws__draws_29_512_050.jsonl","bytes":13172},{"sha256":"cd27000a27203d2fb55b31569c4b991025290db037b4943d30489c0fa2842904","name":"draws__draws_29_512_100.jsonl","bytes":13219},{"sha256":"7ae3e764e180daeee7d420a171f3fa6c71fbbe841780691fb8e1775fd1823b47","name":"draws__draws_29_512_150.jsonl","bytes":13221},{"sha256":"824c31e0aaad0c5d3d85c616b3d8ee3fea7017b0241e6e291ca26b2ecd49c3ea","name":"draws__draws_29_512_200.jsonl","bytes":13217},{"sha256":"ebb5b7e98a9268ad542ea54a10a5078a7483adad5fb20c1c7262db3585dfa351","name":"draws__draws_29_512_250.jsonl","bytes":13216},{"sha256":"dbef139b955fa2c19001fa7c0ef30e702a02766f4faa238ea8a27d59c910fa40","name":"draws__draws_29_512_300.jsonl","bytes":13219},{"sha256":"70c245c340efa0010dc454911b5d1acfa3a4dcda3493587b98bad9e53bfe171a","name":"draws__draws_29_512_350.jsonl","bytes":13220},{"sha256":"889770732def2ad4dbdf61d89751039504d170d6d116fdf7ff1742bc3ec66726","name":"draws__law_19.jsonl","bytes":6242},{"sha256":"dd0581b2c0a719030b57e82ea86130bef74adb0e5dbe73705b0d8aa97305c854","name":"draws__law_23.jsonl","bytes":6408},{"sha256":"e20870effbe41a900a9591e0cbf85342712a278269c1c932c25d914c9db65366","name":"draws__law_29.jsonl","bytes":6671},{"sha256":"0f14e1c06c29497a2500f96a97d0af54020a482d6e77f6bf5566c97b67e9d54f","name":"gate__gate_draws_17_128.jsonl","bytes":97402},{"sha256":"96b1fd98b5bc33756c3f65ed0c5b2fbe0d22ba5e15cae074f04b8cc30250e7f1","name":"gate__gate_draws_17_512.jsonl","bytes":96608},{"sha256":"83418b5d7b96969e3a5b71d75ca453f0e6cb7676a42679a02c00cc48fce13c5d","name":"gate__law_17.jsonl","bytes":5699},{"sha256":"7664ec6507f57ee2c9390b08dd40a77a550815315395688242e4860f104ef788","name":"gate__smoke_draws_29_128.jsonl","bytes":1893},{"sha256":"f0d1768a3dee6904662d2b48c5d68701ca862943aefb57767a756bf3964e6924","name":"make_run_list.py","bytes":1340},{"sha256":"d1d67d43a43ed6457116f4db0b281657e38aca112e5a8be0f9702f7ee9654cbd","name":"neff_draws.py","bytes":11560},{"sha256":"9f7867e7c63581d8f0fa45f6a55ad60faebaeb9454885f545eabab89ceb59aa6","name":"out__analysis.json","bytes":24525},{"sha256":"e38cc6859c274b9dc5301362fe0d18b1f00fd0cbc301ae0a4bbf4215b38de7dd","name":"out__checks.json","bytes":4370},{"sha256":"e2b606c7466c384d6dd2bfec7d3fa6a600d9b98e10410ed6098eccdbfdc295d1","name":"out__tables.md","bytes":3471},{"sha256":"58f50ee2923d6dfeea513212846ebf8b77855d1bc85a6405ca994c50f8acf4ae","name":"run_list.txt","bytes":13092},{"sha256":"c78d14f1d1281ad9db50c5a5b350f29e5310bbc45ce1066d34173184d411ebc1","name":"submit_jobs.sh","bytes":742}],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"report_sha256":"8c678a15344371360010e6f4f6b0078302184d0dbff4f7a23476d1724a93c34e","next_step_sha256":null,"research_authority":{"witness_status":null,"research_status":"pending","scopes":[]},"research_links":[],"duplicates":[],"cited_messages":[]}