{"id":1705,"job_id":3396,"problem_id":1,"lane_id":null,"type":"audit","user_id":1,"model":"deepseek-v4-flash","provider":"deepseek","report_md":"# `research/exponent-control.md`: the ledger's fit basis names the ladder and the fit (#565)\n\nAudit of the served file (`2cd7700f…`, 18004 B, 322 lines). **One line changed**, in the ledger `verdict:`\nof the Q-block: `… stat on 22 trusted terms after control correction …` → `… stat on the 22-term trusted\nladder (fit over the 20 terms p >= 5, p = 2 and 3 excluded as in every fit here) after control correction …`.\n`diff -u`: 1 hunk, +1/−1 line; no other line, number, table or attribution moved.\n\n**Why that wording.** The body of the same file says the G2 rows are \"the 2026-08-21 refit on the 22-term\ntrusted A144311 ladder\" and then \"raw log-log slope over p in [5, 79] (n = 20, p = 2 and 3 excluded as in\nevery fit here)\" — two different counts for two different things, which the old verdict collapsed into \"22\ntrusted terms\". The producer agrees in its own words: re-run here (rc 0, 267 stdout lines), its §5 prints\n`raw fit, 22 terms, p in [5,79], n=20:`, `a = 1.777 +- 0.029 (nominal)`, the matched-width bias\n`mean=1.279 sd=0.052 bias=+0.279` over 45 windows and `1.777 - 0.279 = 1.498` — so 1.50 is 1.498 rounded.\nThe clause also matches `paper/two-class-jacobsthal.md` §4 (\"20 terms with $x \\ge 5$ of the 22-term ladder\n($p = 2, 3$ excluded, as in every fit of that note) returns $1.50 \\pm 0.05$\"), which is the cross-document\nconsistency the finding asks for.\n\n**No re-embedding.** `fences = code-sha256 = out-sha256 = 0` on the served text and on the revision; the change\nis prose inside the ledger comment, and the file's numbers are all the producer's, unchanged. The producer was\nrun anyway because the file names it as the reproduction path, and every figure the changed sentence rests on\nreproduced (`1.777`, `+0.279`, `1.498`; stdout sha256 633811cfbc4bcd4bfe91eaa070ad4cd13cb8c4fb26c032975fc5e2e4a61fb738).\n\n**Left open, not this finding.** `QUESTIONS.md` is generated from these ledger blocks and is not hand-edited\n(the corpus rule), so its row still shows the old wording until the next index run; the ledger block itself is\nthe source and is now correct. Finding **#21** on the same file (\"conjectured exponent 1\", the proven bracket\n[1, 2]) is open and outside this job's brief; it is left for the job that is assigned it.\n\n**Evidence.** Served file fetched 2026-09-25 and re-GET immediately before the return (sha unchanged); finding\n#565 with `content_sha` = the served sha; return #1261 for the review; return on job #78 is the origin of the\n#21 note cited above. Uploads: the revision and `evidence-3396.md`.\n\n50 of @Benjaminsen's returns wait for a verdict (this job's brief).\n","patch":null,"cpu_hours":0,"hashes":{},"author_rung":null,"status":"pending","final_rung":null,"created_at":"2026-09-25T13:34:15.759Z","repo_url":null,"commit":null,"cites":{"returns":[1261]},"tokens":{"log":"custom","input":0,"models":{"deepseek-v4-flash":0},"output":0,"source":"none","entries":0,"cache_read":0,"cache_write":0,"observed_models":["deepseek-v4-flash"]},"paper_slug":null,"revision_path":"research/exponent-control.md","revision_sha":"297ea0b208ba809376bd8e09a383f1735b0b56e731e17163967d8d9bce0a0c39","recipe_md":"1. GET /projects/twin-primes/docs/research/exponent-control.md -> sha256 2cd7700f... (18004 B, 322 lines).\n2. Findings: GET /projects/twin-primes/findings?path=research/exponent-control.md -> #565 (and #21, open,\n   outside this brief).\n3. Comparison texts: the file's own line 192 and 186-187; GET to docs/paper/two-class-jacobsthal.md §4.\n4. Edit: one anchored replacement in the ledger verdict, asserted to occur once, with the ledger block count\n   and the body's `n = 20, p = 2 and 3 excluded as in every fit here` asserted intact.\n5. Reproduction: GET docs/research/exponent-control.js, `node exponent-control.js` (rc 0, ~1 s, 267 lines)\n   and grep the §5 block: `raw fit, 22 terms, p in [5,79], n=20:`, 1.777 +- 0.029, bias=+0.279, 1.498.\n6. POST /files (revision, evidence-3396.md), then POST /projects/twin-primes/result with base = served sha.","verification":null,"target":null,"finding":null,"human_md":null,"provisional":false,"effects_applied_at":null,"effort":null,"also_fix":null,"transcript_omitted":{"share":0,"omitted":0,"outputs":0},"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":"2026-09-25T13:34:15.759Z","department_id":"dept_0e793a31e299699dfaaa6fee","run_id":"run_d8f7b9a37c651d807bfda54e","triage_lead":null,"revision_base_sha":"2cd7700fca1c000b514382d338e0e52db33f0d6438fe47bba63dceebb81e6752","integration":null,"resolves":[565],"handle":"Benjaminsen","job_brief":"A reviewer found a defect in the served file `research/exponent-control.md` while reviewing return #1261 (review #350 by @Benjaminsen), recorded as finding #565. Fix it; do not redo the work it belongs to.\n\nWhat the reviewer said:\n> Ledger verdict (line 8) says \"MEASURED 1.50 +/- 0.05 stat on 22 trusted terms\", but the fit it reports (line 192) is \"slope over p in [5, 79] (n = 20, p = 2 and 3 excluded as in every fit here)\". Say \"on the 22-term trusted ladder (fit over the 20 terms p >= 5)\" so the QUESTIONS row matches the body and paper/two-class-jacobsthal.md §4.\n\nFetch the current file (GET <project base>/docs/research/exponent-control.md), make the change, check it still runs and that its stdout reproduces byte for byte elsewhere (progress, timing and rates go to stderr; paths relative to the repository), upload the revised file (POST /files) and return as this job with `\"revision\": { \"path\": \"research/exponent-control.md\", \"file\": \"<sha256 of the revised file>\" }`, the sha in `files`, a one-line report of what changed and why, and `\"cites\": { \"returns\": [1261] }`. If the file's embedded hashes depend on the change, re-embed them and say so. Send `\"revision\": { …, \"base\": \"<X-Content-SHA256 of the text you edited>\" }` so a later change to the file is caught rather than overwritten, and list the findings your revision answers in `\"resolves\": [<finding ids>]` (GET <project base>/findings?path=research/exponent-control.md lists the open ones). Accepted, the revision becomes the served version and closes the findings it answered; a finding it leaves open goes to the next fix job.","review_deferred":false,"in_triage":true,"triage":[],"verification_runs":[],"verification_state":null,"verification_summary":null,"canonical_return":null,"review_history":[],"dependencies":[],"research_url":null,"transcript_url":"/projects/twin-primes/return/1705/transcript","files":[{"sha256":"297ea0b208ba809376bd8e09a383f1735b0b56e731e17163967d8d9bce0a0c39","name":"research-exponent-control.md","bytes":18088},{"sha256":"438cdeba2d76b99061e3748303d2b0a7bb6e6753a052b50545b2f346889a86fa","name":"evidence-3396.md","bytes":1826}],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"duplicates":[],"cited_messages":[]}