Back to dashboard
m23-b4-5a4M235A^4B4claimed

Check Frobenius trace consistency for positive candidates

Check a surviving chunk against Frobenius-style consistency filters before the route spends more orbit time.

Verified chunk progress
29%
2 completed · 5 remaining

What the computation does

  1. 1Load M23 in GAP or the configured computation backend.
  2. 2Load the Atlas class 5A.
  3. 3Fix the first branch-cycle representative a.
  4. 4Enumerate (b, c) in the assigned chunk rectangle.
  5. 5Set d = (a*b*c)^-1.
  6. 6Keep tuples where d is in class 5A.
  7. 7Check that the generated group is M23.
  8. 8Compute the inner braid orbit under Hurwitz moves.
  9. 9Canonicalize using the fixed first representative and centralizer quotient.
  10. 10Stop when the orbit closes or exceeds threshold + 1.
  11. 11Emit a structured JSON packet with hashes and scope caveats.

Use ChatGPT Pro

Copy one prompt. Upload a redacted transcript as review credit only.

Run in Colab

Download the notebook, open it in Colab, and click Run All.

Download notebook

Cloud workspace

Use a preconfigured web IDE when the public repo is connected.

Open Codespaces

Run locally

Copy one command. It prints the task before doing any work.

Safety task card

Public inputs

  • Candidate hashes.
  • Trace modulus list.
  • Checker version.

Expected outputs

  • Pass/fail counts.
  • Failing witness hashes.
  • Runtime.

Data submitted

  • Result JSON.
  • Input and output hashes.
  • Logs hash.

Data never requested

  • Secrets.
  • Private files.
  • AI keys.

Prohibited actions

  • Do not submit unreviewed failing witnesses.

AI scope

  • AI may explain trace constraints but must not be trusted as verification.

Terminal worker

No AI provider is called. The worker prints the task card and scope caveat before computation.

node public/m23/worker.mjs --work-unit m23-frob-0427 --i-understand-task

Local AI or A2A

Keys stay local. The command stops unless the explicit AI acknowledgement is present.

M23_AI_ENDPOINT=http://localhost:11434/v1/chat/completions node public/m23/worker.mjs --work-unit m23-frob-0427 --run-ai --i-understand-ai-task
node public/m23/worker.mjs --work-unit m23-frob-0427 --agent a2a --endpoint http://localhost:41241 --i-understand-ai-task

Upload result packet

Paste the JSON packet printed by the notebook or worker. It will be pending until another person reproduces it.

ChatGPT Pro review

Copy the prompt into ChatGPT Pro, redact the transcript, then upload it as review context. This earns review credit only.

I am contributing to m23isgalois.com.
Please review this public mathematical compute task. Do not ask for secrets, keys, cookies, private files, or account data.
Work unit: m23-frob-0427
Title: Check Frobenius trace consistency for positive candidates
Route: m23-b4-5a4
Task card hash: sha256:0427f5a45a4b4b4b0427f5a45a4b4b4b0427f5a45a4b4b4b0427f5a45a4b4b4b
Task: Check a surviving chunk against Frobenius-style consistency filters before the route spends more orbit time.
Expected packet schema: { caseId, routeId, workUnitId, status, orbitSize, threshold, tupleSummary, inputHashes, outputHashes, notes }
Scope caveat: This result covers only the assigned chunk and does not prove the M23/Q question by itself.
Your job: explain what this task does, list what evidence a valid packet would provide, and list what it would not prove.
Return concise review notes. This review is not a proof and does not replace independent reproduction.

Submissions

No submissions yet.

Reviews

No ChatGPT Pro reviews yet.

Verifications

No reproduction record yet.