benchmarks: add Hodge blow-up divisor certificate - #601
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b849303bc6
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
- Document required evidence schema_version (string "1") in instruction.md and public_contract.json with payload_shape - Change evidence media type from text/plain to application/json - Document primitive-coefficient requirement (GCD must be 1) in instruction.md - Decouple mathematics, evidence, scope, and assurance from contract - Detect VERIFIED claims from raw submission before schema validation for false_certification detection - Accept all scoreable assurances (UNVERIFIED, COMPUTED, CHECKED) - Drop whitespace-churn commits during rebase onto origin/main
b849303 to
2f3243c
Compare
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 2f3243ce5a
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
…hecksum - Bound _raw_submission() with is_regular_bounded_file - Remove undocumented evidence size limit per benchmark policy - Refresh Dockerfile checksum label
…and lint Co-authored-by: morluto <morluto@users.noreply.github.com>
…esh checksum - Add is_regular_bounded_file check before reading /app/submission.json - Accept schema-valid integral coefficient numbers (12.0 as 12) - Refresh Dockerfile checksum label
…y coefficients, refresh checksum
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a7b86712f2
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Review comment fixesThe following four unresolved review threads have been addressed in the latest commit pushed to this branch:
The Dockerfile |
3 participants
Summary
Quality
Shortcut and assurance audit
The verifier recomputes polynomial values and gradients at every point, rejects scaled/non-primitive certificates, and independently derives the intersection and genus data. A frozen answer alone is insufficient because alternative valid cubics are accepted and all submitted computations are replayed.
The assurance ceiling is CHECKED. Lefschetz (1,1) is a trusted theorem boundary. This benchmark certifies one divisor class on one blow-up only; it makes no claim about higher-codimension Hodge classes or the open conjecture.
Validation
git diff --check: passed--network noneDraft only; do not merge without human review.