{"id":5524,"job_id":11192,"problem_id":1,"lane_id":null,"type":"measure","user_id":103,"model":"claude-opus-5-5","provider":"anthropic","report_md":"# Route 295: class-2 enrichment at 23# -> 29# is E = 1.2193, E/E_ans = 0.680 (success clause fires); the decline sits in u's arrangement channel, within-class E_w rises to 1.544\n\n## Result\n\nMeasured, exact, one more rung (finite scope). At q = 23#, Q = 29# (N_A = 214,708,725 gaps) the\nroute's enrichment is **E = 1.2193** against **E_ans = 1.7924**, so **E/E_ans = 0.6802**. E lands inside\nthe pre-registered [1.10, 1.35] and under the 0.75 bar, so **the route's success clause fires as\nwritten**. The pre-registered secondary reading also fires, and it changes what that clause means.\nThe route's u = V/(V + |rho_1|) is not the class-2 share of the centred variance. An exact identity\n(section 2) shows it carries the arrangement channel A through the inversion. A moved from +2.45e-2\nto -1.16e-2 between 17# and 29#, and that movement produces the decline in E. The within-class share\nitself, E_w = u_w/(k/N), **rises**: 1.1474, 1.3499, 1.4166, **1.5442** at 17#, 19#, 23#, 29#. The\nclass-2 within-class variance stays within 2-11 % of the moment value 2 var_q (1 + rho_1(q)) at every\nrung (ratio 0.899, 0.973, 0.951, **0.981**). E_ans's shortfall comes mostly from its C_0 composition,\n1 + (k/N)(1 + 2 rho_1(q)), which leaves out the between-class mean shift: it gives var_Q/var_q =\n1.065 at 29#, and the measured value is 1.213. After the run (post hoc, not pre-registered), a\ncomposition built from the same q-word moments plus the census (block algebra) with the mean shift\nincluded reproduces var_Q/var_q to 0.5 % or better. It predicts E_w with ratios 0.894, 0.969, 0.951,\n0.982, and V to +14.1 %, +3.8 %, +6.0 %, +2.1 %. That functional is **not certified at the carrier\nMonte-Carlo bar** of the original step (at 19# it misses V by 3.1e-4, 13.6 times #4258's SE of\n2.3e-5). One rung says nothing about \"any larger rung\" in either direction. What the run does show\nis that the decline the route tests is a property of its own u, not of the dispersion of merged\npairs.\n\n## 1. What was run\n\n- `class2.c`: a segmented sieve of A(x#) over odd n (x = 7..29), and the p-block lift q# -> Q#. The\n  lift builds the class word (merge multiplicity m), the Q-word stats and the class-2 accumulators,\n  all as uint64 integers. At 29# it also writes the word to a scratch directory outside git\n  (3 bytes per gap: uint16 gap, uint8 m; 644,126,175 bytes, sha256\n  `39d303df08c7bb5f4938d3aaa456d86584ccb8798cf0b352c5e1708358663f9e`). That is packed, not\n  the 1.7 GB int64 layout the route estimated.\n- `analyse.py`: exact rationals (`fractions`), gates G1-G7, the channel table.\n- `check.py`: pure Python, no code shared with the two files above. It uses its own bytearray\n  sieve, builds the class word by testing each lifted residue directly, gets Delta_2 by rewriting\n  the word (scaled by k) and recomputing, and gets A from seeded exchangeable draws. This checks\n  the arithmetic. It is not a separate derivation: both paths implement the same definitions.\n- `composition.py`: the post-hoc side check of section 4.\n- `PREREGISTRATION.md`: committed in 26109e7, before any 29# output existed. The 29# outputs are in\n  the next commit.\n\n## 2. Definitions and the identity\n\nNotation: rho_k = (N S_k - P^2)/(N S_0 - P^2), S_k = sum g_i g_{i+k} cyclic, P = sum g = Q#. Class 2\nis the Q-gaps with m = 2; there are k of them, with sum T1 and sum of squares T2, and\nw = T2 - T1^2/k. The neutralisation replaces each class-2 gap by T1/k (#4191). That preserves P,\ngives S_0' = S_0 - w, and S_1' = S_1 - P22 - X12 + n22 (T1/k)^2 + (T1/k) Y12. Here P22 and n22 are\nthe products and count of adjacent class-2 pairs; X12 and Y12 are, over adjacent (class 2, other)\npairs, the sum of products and the sum of the other gap.\n\nWrite C_0 = N S_0 - P^2, A_cm = N (S_1 - S_1')/C_0 (the arrangement channel at fixed C_0) and\nu_w = N w / C_0. Then\n\n    rho_1' = (N S_1' - P^2)/(C_0 - N w) = (rho_1 - A_cm)/(1 - u_w),\n    Delta_2 - A_cm = rho_1 - A_cm - (rho_1 - A_cm)/(1 - u_w) = -(rho_1 - A_cm) u_w/(1 - u_w).\n\nProof: substitute N S_1' - P^2 = (N S_1 - P^2) - N(S_1 - S_1') and divide by C_0. QED. Gate G6\nchecks it to exact zero at every rung.\n\n#4258's A is the exchangeable carrier's mean error with the sign flipped, at fixed C_0. Its closed\nform differs from A_cm only in the adjacent class-2 term, E[x_i x_j] = (T1^2 - T2)/(k(k-1)) in\nplace of (T1/k)^2. So A_cm - A_MV = -u_w n22/(k(k-1)), which is -3.4e-6, -1.7e-7, -1.1e-8 and\n-3.6e-10 at 17#, 19#, 23#, 29#. The route sets V = Delta_2 - A and u = V/(V + |rho_1|), which\nis the inverse of V = |rho_1| u/(1-u) (#5492 section 3). That inverse treats A as normalised by\nC_0 - w, but #4258's A is normalised by C_0. The exact relation is\n\n    u/(1-u) = [u_w/(1-u_w)] * |rho_1 - A_cm| / |rho_1|,\n\nso u = u_w only when A = 0. At 29#, |rho_1 - A|/|rho_1| = 0.769.\n\n## 3. The table (exact; `out/analyse-17-29.txt`, gates 0 FAIL)\n\n| Q | k | k/N | rho_1(Q) | Delta_2 | A | V | E | E_ans | E/E_ans | E_w |\n|---|---|---|---|---|---|---|---|---|---|---|\n| 17# | 2,826 | 0.126869 | -0.0397484 | +3.5426e-2 | +2.4486e-2 | +1.0940e-2 | 1.7012 | 1.6880 | 1.0078 | 1.1474 |\n| 19# | 42,374 | 0.111901 | -0.0420612 | +1.2913e-2 | +4.6088e-3 | +8.3040e-3 | 1.4734 | 1.7412 | 0.8462 | 1.3499 |\n| 23# | 733,672 | 0.092261 | -0.0452721 | +3.3029e-4 | -5.6298e-3 | +5.9601e-3 | 1.2609 | 1.7666 | 0.7138 | 1.4166 |\n| **29#** | **15,416,706** | **0.071803** | **-0.0503153** | **-6.7760e-3** | **-1.1604e-2** | **+4.8276e-3** | **1.2193** | **1.7924** | **0.6802** | **1.5442** |\n\nAt 29#, u = 0.087548 and u_w = 0.110880. The census at 23# -> 29# is m = 1/2/3/4 =\n199,048,197 / 15,416,706 / 243,822 / 0, so sum_{m>=3}(m-2)#m = 243,822 (no published value to\ngate against). The class-2 footprint Delta_2 changes sign between 23# and 29#.\n\nMechanism quantities (`out/analyse-17-29.txt`, second table):\n\n| Q | (w/k) / (2 var_q (1+rho_1(q))) | var_Q/var_q | E_ans composition | corr(a, b) in class 2 | T_adj | cross |\n|---|---|---|---|---|---|---|\n| 17# | 0.89903 | 1.46951 | 1.11108 | -0.0932 | +1.975e-3 | +2.251e-2 |\n| 19# | 0.97303 | 1.38431 | 1.10300 | -0.0562 | +2.777e-4 | +4.331e-3 |\n| 23# | 0.95125 | 1.28650 | 1.08450 | -0.0487 | -2.205e-3 | -3.425e-3 |\n| 29# | 0.98119 | 1.21325 | 1.06530 | -0.0376 | -1.695e-3 | -9.909e-3 |\n\nHere a and b are the two q-gaps a merged pair is made of. T_adj and cross split A_cm (in units of\nC_0), and they reproduce #4258's T_adj of +2.06e-3, +2.79e-4, -2.20e-3 to the printed digits. At\n17#, E's agreement with E_ans comes from two offsetting differences: E/E_w = 1.483 there, and\nE_ans/E_shift = 1.316.\n\nThe arrangement channel's leading-order value, A_lo = (k/N) 2(rho_1(q) + rho_2(q))/(1 + (k/N)(1 + 2\nrho_1(q))), is +1.31e-2, -3.76e-3, -1.47e-2, -1.61e-2 at 17#, 19#, 23#, 29#. The exact A is\n+2.45e-2, +4.61e-3, -5.63e-3, -1.16e-2. A_lo has the wrong sign at 19# and misses by a factor\n0.5-2.6 elsewhere, so this leading-order form does not carry the arrangement channel at any of the\nfour rungs. That covers this functional only, not every functional of the moments.\n\n## 4. Pre-registered decision, and the post-hoc side check\n\n- **Primary (route 295's clauses).** E(29#) = 1.2193 is in [1.10, 1.35], and E/E_ans = 0.6802 is\n  at most 0.75: **success fires**. The failure clause (E >= 1.5) does not.\n- **Secondary (registered blind expectations).** E_w(29#) = 1.5442, in [1.35, 1.55]. The\n  dispersion ratio is 0.981, in [0.90, 1.00]. var_Q/var_q = 1.213 stays above the composition's\n  1.065. All three came in as expected.\n- **Reading rule (registered).** The primary fires and E_w does not fall. So the decline is a\n  property of the route's u, which carries A through the inversion of section 2. It is not a\n  property of the class-2 within-class variance share.\n- **Post-hoc side check, not pre-registered (`composition.py`, `out/composition-17-29.txt`).** A\n  class-m Q-gap is taken as the sum of m adjacent q-gaps, with mean m mu_q and variance\n  var_q (m + 2 sum_{d<m} (m-d) rho_d(q)). The census gives the fractions f_m, and mu_Q = Q#/N is\n  exact. That gives var_Q/var_q = 1.4622, 1.3789, 1.2857, 1.2144, against the measured 1.4695,\n  1.3843, 1.2865, 1.2133, and E_shift = 2(1 + rho_1(q))/(that) = 1.2827, 1.3928, 1.4901, 1.5724.\n  E_w/E_shift = 0.894, 0.969, 0.951, 0.982. Through the identity, with the exact A, V is predicted to\n  +14.1 %, +3.8 %, +6.0 %, +2.1 % (+9.9e-5 at 29#). This functional reads only q-word moments plus\n  the census. It fails the original step's carrier bar at 19# (3.1e-4 against SE 2.3e-5), so it is\n  **not certified** there. It was formed after the 29# numbers were seen.\n\nWhat the route's success clause goes on to conclude is \"the closure-compatible repair of the class-2\nC0 channel is closed at and above the crossing rung\". This run does **not** support that. First,\nnothing about rungs above 29# follows from 17#..29#. Second, the quantity that declines is a mix of\nthe C0 and C1 channels. Read as u_w, the C0 channel is predicted by a q-moment-plus-census composition at all four\nrungs, E_w within 2-11 % and V within 2-14 %, with no trend away. What the measured\nleading-order forms do not carry, at every rung, is the arrangement channel A: the cross-coupling\n#4258 identified.\n\nBearing on #5492: its reading that \"the merged pairs are relatively less dispersed than two\nadjacent q-gaps, increasingly so with p\" is not borne out at 17#..29#. The measured dispersion\nratio is 0.899, 0.973, 0.951, 0.981, with no decreasing trend. #5492 flagged the C_0 composition\ninside E_ans as unverified (its recipe, claim 3). Against a measured C_0, that composition is short\nby 12-24 % in var_Q/var_q (24 %, 20 %, 16 %, 12 % at 17#, 19#, 23#, 29#), and that is what this run finds. #5492's verdict on the\nearlier step (no functional meets the 2.3e-5 carrier bar at 17# -> 19#) is not contradicted: the\npost-hoc functional above does not meet that bar either.\n\n## 5. Gates (all PASS at 17#..29#; `out/analyse-17-29.txt`)\n\n- G1: N_A = prod(p-2) and sum g = x# at 11#..29#.\n- G2: rho_1..4 match #2426/#2492 to < 5e-7 at 11#..23#, and #2428's 29# row\n  (-0.050315, -0.085892, -0.134505, -0.076399) to 4.4e-7.\n- G3: the lift and the direct segmented sieve agree on N, P, S_0..S_4 for both words at every\n  transition, 29# included.\n- G4: sum census = N_A(Q#) and sum m #m = p N_A(q#). sum_{m>=3}(m-2)#m = 2N0+N2 = 72, 1088, 11870\n  (#4191). The census equals #4191's at m >= 3 and differs by m=1: -4, m=2: +4. That is #4258's\n  documented convention, which this run uses. #4191's own boundary convention was not\n  reconstructed.\n- G5: Delta_2 matches #4191 within 1e-4. The exact -A matches #4258's MV mean within 3 SE\n  (z = +0.24, -0.03, +0.76). A matches #4258's printed arrangement channel.\n- G6: the identity of section 2, exact.\n- G7: #5492's formula on published inputs gives 1.6987, 1.4715, 1.2611, and on exact inputs 1.7012,\n  1.4734, 1.2609, against the printed 1.699, 1.471, 1.261. E_ans agrees to 1e-3.\n- `check.py`: Delta_2 is identical at 17#, 19#, 23#. A by draws: 2.510e-2 +- 5.1e-4 (40 draws),\n  4.734e-3 +- 1.1e-4 (40), -5.623e-3 +- 5.3e-5 (10), within 1.2 SE of the closed form. u_w and E_w\n  are identical.\n- Stored 29# word: re-read from bytes, N = 214,708,725, sum g = 29#, census identical\n  (`out/word-29-check.txt`).\n- 11#->13# is reported only. There the 4-gap convention is 4 of 258 class-2 gaps; #5492 left it\n  out as well.\n\n## 6. CPU time (measured, getrusage or /usr/bin/time, one core, `nice -n 10`)\n\n| step | CPU s |\n|---|---|\n| direct sieves 7#..23# | 0.18 total (23#: 0.155) |\n| lifts 7#->11# .. 19#->23# | 0.08 total (19#->23#: 0.062) |\n| **direct segmented sieve 29#** | **4.681** |\n| **lift 23#->29# + class word + 644 MB write** | **1.854** (A(23#) sieve 0.150) |\n| analyse.py, all rungs | 0.03 |\n| composition.py | under 0.1 (not separately timed) |\n| word file re-read check | 3.40 (10.9 wall) |\n| check.py 17# / 19# / 23# | 0.14 / 2.09 / 9.83 (42.7 wall) |\n| **total** | **about 22 CPU s** |\n\nThe route's estimate was 0.5 CPU h. The Studio runs happened 07:53Z-08:00Z on 2026-10-23, during\nthe integrate-round2 exclusive hold (07:42Z-08:42Z). Each run was under HEAD-NOTES item 8's\none-minute threshold, at most one labelled worker at a time (`r295-direct29`, `r295-lift29`,\n`r295-check23`). None is still running. Nothing runs in the 09:00Z-10:00Z window.\n\n## 7. Not run\n\n- Any Q above 29# (out of scope). In particular, no pre-registered test of the post-hoc composition\n  functional at 31#, which is the obvious next step.\n- No closed form for the arrangement channel beyond the leading-order A_lo. No insertion of any term\n  into #2492's closure. No rho_2 channel.\n- No exchangeable draws at 29#: A there is the exact closed form only. check.py was run at 17#, 19#,\n  23#, not at 29#.\n- The 1.7 GB int64 word layout: the word is stored packed at 644 MB, in a scratch directory outside\n  the repo.\n- #4191's 4-gap census convention was not reconstructed. The 4 gaps change k/N by 5e-6 relative at\n  23#.\n- No outside expert round and no filing (the head's call).\n\n## What this rests on\n\nThe exact integer runs above. Definitions from #4191 (neutralisation, census, identity), #4258 (the\nexchangeable carrier, the arrangement channel at fixed C_0, T_adj), #2426/#2492/#2428 (the rho\nladder), and #5492/route 295's next_step (E, E_ans, the clauses). The identity in section 2 is\nproved above. Every other number is a finite exact computation, labelled measured; the side check of\nsection 4 is labelled post hoc.\n\n## Reproduce\n\n```\nSCR=$HOME/scratch/r295-class2-29; mkdir -p \"$SCR\"\nclang -O2 -o \"$SCR/class2\" class2.c\nfor x in 11 13 17 19 23 29; do nice -n 10 \"$SCR/class2\" direct $x > out/direct-$x.json; done\nnice -n 10 \"$SCR/class2\" lift 19 23 > out/lift-19-23.json        # likewise 11 13, 13 17, 17 19\nnice -n 10 \"$SCR/class2\" lift 23 29 \"$SCR/word-29.bin\" > out/lift-23-29.json\npython3 analyse.py out --rungs 17,19,23,29                          # gates 0 FAIL, exit 0\npython3 composition.py out 17 19 23 29\nnice -n 10 python3 check.py 23 --draws 10                           # about 45 s wall\n```\n\n## Filing note\n\n- Route **295**, scope key **`r295-class2-29`**.\n- Links: **#5492 extends** (executes its next_step at 23# -> 29#; also bears on its\n  dispersion-mechanism reading, section 4); **#4191 bears_on** (census, neutralisation, Delta_2 at\n  29#); **#4258 bears_on** (A at 29#, and the fixed-C_0 normalisation that the u inversion mixes).\n- Outcome for the head to choose: the route's success clause fires as written (E = 1.2193,\n  E/E_ans = 0.6802). The registered reading rule also fires: the decline sits in u's A-dependence,\n  and E_w rises to 1.5442. The next step that would decide something: a pre-registered test at 31#\n  of the q-moment-plus-census composition for u_w and V (not certified at the 2.3e-5 bar at 19#),\n  and a closed form for A.\n- Cheapest decisive check: E at 23# rerun from the committed script,\n  `clang -O2 -o class2 class2.c && ./class2 direct 23 > out/direct-23.json && ./class2 lift 19 23 >\n  out/lift-19-23.json && python3 analyse.py out --rungs 23`. That takes under a second and should\n  print E = 1.2609, E_w = 1.4166, gates 0 FAIL.\n\n\n## Filing\n\n`out/x` files are attached as `out__x` (the server takes no subdirectories); restore them under `out/` before running `analyse.py`.\nThe two `.time` files are attached with a `.txt` suffix added.\n","patch":null,"cpu_hours":0.006,"hashes":{},"author_rung":"measured","status":"pending","final_rung":null,"created_at":"2026-10-23T08:06:37.473Z","repo_url":null,"commit":null,"cites":{"returns":[5492,4191,4258,2492,2428]},"tokens":{"log":"claude-code","input":10,"models":{"claude-opus-5-5":1648},"output":1648,"source":"claude-jsonl","entries":5,"cache_read":740633,"cache_write":5793,"observed_models":["claude-opus-5-5"]},"paper_slug":null,"revision_path":null,"revision_sha":null,"recipe_md":"Cheapest decisive check, E at 23# (under a second): `clang -O2 -o class2 class2.c && ./class2 direct 23 > out/direct-23.json && ./class2 lift 19 23 > out/lift-19-23.json && python3 analyse.py out --rungs 23` prints E = 1.2609, E_w = 1.4166, gates 0 FAIL. The 29# rung: `./class2 lift 23 29` then `analyse.py out --rungs 29`; whole run about 22 CPU s.","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":5},"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":"2026-10-23T08:06:37.473Z","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 295's next step: at q = 23#, Q = 29# (214,708,725 gaps), build the class-2 word exactly, measure Delta_2 and A at fixed C_0 with #4191's and #4258's neutralisations, report V, u, E against E_ans, the arrangement channel's value, and test the pre-registered E(29#) in [1.10, 1.35].\n\nWhy this step: The measured enrichment fell 1.699, 1.471, 1.261 at 17#, 19#, 23# against a moment prediction near 1.79; one more rung decides whether the shortfall is a trend.\n\nStop when: The return is accepted with E at 29# from an exact class-2 word that passes the census and rho-ladder gates, the clause verdict stated, and the channel decomposition reported; 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/5524/transcript","files":[{"sha256":"d55750ab328e4e525c254f4997e788564cea0cd6f2292e4cf3b596dbcc5101f1","name":"PREREGISTRATION.md","bytes":6658},{"sha256":"fd2dd5dcc1fc02c8d9677c96c57c5a0a46672fb289f701116e0f86b2b0f37e6c","name":"RESULT.md","bytes":14987},{"sha256":"c7646d5e3ae3886bd0979cbae97dfa0d425ff2baa1914ffcd910e7f932850d1d","name":"analyse.py","bytes":12575},{"sha256":"f12c07f64c46b0a05b71482bc11699169f0059135fab93395cffd2fe35291736","name":"check.py","bytes":4235},{"sha256":"22181a25d732d39c5a10de91ff107fe34300d562913fbf625ce979737e4ee117","name":"class2.c","bytes":9723},{"sha256":"a823a87669f674ff47914ed169af74235a62e6fa3d1aa89ea7376214959b0fbc","name":"composition.py","bytes":2432},{"sha256":"d1c312b1d9266dfeaa79f4920f3e5a61d89c195cb9996a3b3b75da49769bb0d8","name":"out__analyse-13-23.txt","bytes":5907},{"sha256":"302732af76ca3eab791b52904a4ed3b0f543b9387f423c028783f4353e0912c6","name":"out__analyse-17-29.time.txt","bytes":29},{"sha256":"d845ce1eda51584cc55917bf64005dead3f3fb5095cb355efd0997c5e97d1383","name":"out__analyse-17-29.txt","bytes":5374},{"sha256":"9f3f36b78b3c6e3d062f33c49c7d22dd6986d8e1f39ee112ecf1f47e3a2541e4","name":"out__channels-13-23.json","bytes":3829},{"sha256":"911453c99a33970b2efbcaca5e842a06211bc5c2d048c446b0674545e1ea3a05","name":"out__channels-17-29.json","bytes":3853},{"sha256":"9257af77bb4ec2661f9fc269b6854719d9eef189ac73e4a29f1bf1e8b5d38f3b","name":"out__check-17.txt","bytes":300},{"sha256":"d4dc60bf7e75f9d6e0fb4f8f578054b9cc5dbae041d8b69bb7bf4e288e7fc1da","name":"out__check-19.txt","bytes":302},{"sha256":"1a5691770966efdd031e52d6e952ab7b8e3d8abf3150973d5765173718a77414","name":"out__check-23.time.txt","bytes":30},{"sha256":"c3131e1563ab240b19330a99d398c26a0901338c23e7cec0fc9c9a20ecc0438e","name":"out__check-23.txt","bytes":304},{"sha256":"18ef2b7c955c105833abd61dc98c8841928b6dab0954232e3e93a0e38f3475ff","name":"out__composition-17-29.txt","bytes":548},{"sha256":"87b7d932e9facbb7f61bd1d72abf0e12f5eca401497b115f7ed1f19f30e92ffc","name":"out__direct-11.json","bytes":184},{"sha256":"186a6b392a9ccd0a507928dba0e230dae3f6750272b3bf200e84cb042ff5c8fd","name":"out__direct-13.json","bytes":192},{"sha256":"fa3d04b997027ff04d8567cff98453fc14a4054aab7f580d88b14baaace3b8a5","name":"out__direct-17.json","bytes":206},{"sha256":"f21358d2e7b48882cf6930c2162608fbf7103f3523a3a4002d492cf5df596595","name":"out__direct-19.json","bytes":214},{"sha256":"f755134a1a72cdba00f804878b0129841b3e5ccf97b2bebe48060ff934a3480e","name":"out__direct-23.json","bytes":224},{"sha256":"f36e074214b4d7e37426908632536698a492c172791c023307c6207f8fd21e13","name":"out__direct-29.json","bytes":238},{"sha256":"364f7a19aa7327aaa6d377897dcb134e8a67462d112468cb971e599b8d45c89c","name":"out__direct-7.json","bytes":175},{"sha256":"c5f3e4c6920179537b29e4152f3282cf9f4fe8232c6f1b38c9db501a81c2f459","name":"out__lift-11-13.json","bytes":716},{"sha256":"2181defcf64c7b5c7472225f26b162451bf52f4c43e9a366f363cff8c2499309","name":"out__lift-13-17.json","bytes":762},{"sha256":"534c316aea4d0ca46bb5952652165833eb5f41398d7fde84877300705a1e28cf","name":"out__lift-17-19.json","bytes":804},{"sha256":"961a92e3bf2a429d374095d8d34002939cbfb24c959ccf437fc36dbba9f241bd","name":"out__lift-19-23.json","bytes":846},{"sha256":"14778abb60b68a9270a8652bb69a4362819068c323b6ae3f88b61f594df3f929","name":"out__lift-23-29.json","bytes":894},{"sha256":"fb85e9d9aa1e9b4989cde40f4898e4f692e2a0a42e6d1fccbfb43ecc9f02923d","name":"out__lift-7-11.json","bytes":683},{"sha256":"8d2a76657e8111e9a054eec06ec665ab41dbfb4751e88f41ac87afb41fa1deb4","name":"out__word-29-check.txt","bytes":205}],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"report_sha256":"54895faa056e6f16925767d8bf7e092394ebead6dad126e31210e13ff0dc27d9","next_step_sha256":null,"research_authority":{"witness_status":null,"research_status":"pending","scopes":[]},"research_links":[],"duplicates":[],"cited_messages":[]}