Add chrisrcole submission entry - #90
Open
chrisrcole wants to merge 3 commits into
Open
Conversation
Cole/Gibbs collaborative MIB doc challenge submission: predictions.jsonl, technical memo, and a link to the public Dockerfile-based solution repo. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
… run 5,000 valid rows, 0 missing case ids against the official manifest. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AB9ijdCBq5m6ueikshJ8Ks
The first PR was opened at the human submission deadline, before this container's own run had reached the contract's 30,000s / 6s-per-PDF time budget. This commit swaps in the prediction file exactly as it stood at that budget cutoff (container start + 30,000s) -- the score this code would actually receive under the official runtime contract, not a later, unconstrained-time result that the real contract would never have allowed. Run receipt: wall 29,890s for 5,000 PDFs = 5.98 s/PDF (100% of budget); pass 1 completed all 5,000; the governed refinement pass covered 757 worklist cases before the deadline stop; container exited 0 on its own. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AB9ijdCBq5m6ueikshJ8Ks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds
submissions/chrisrcole/per the challenge's submission format:predictions.jsonl— validation-set predictionsMEMO.md— technical memo (approach, attribution, failure modes, what's next)SUBMISSION.md— link to the public Dockerfile-based solution repositoryApproach
Builds on Tyler Gibbs's audited MIT-licensed baseline (full attribution
chain in the solution repo's
ATTRIBUTION.md/NOTICE.md) with additionalnote-verdict recovery, page classification, and time-governed runtime
refinement layered on top. Full technical detail, measured results, and
known failure modes are in
MEMO.md.Checklist
predictions.jsonlvalidates againstscripts/validate_submission.pySUBMISSION.mdlinks a public repository containing aDockerfilesubmissions/chrisrcole/Timing — submitted after the deadline
This PR is being opened after the 11:59 PM PT submission deadline; the entry
was staged before the cutoff but intentionally held. The predictions file is
the run's state at exactly its official 30,000 s / 6 s-per-PDF budget — not a
longer, unconstrained-time result: wall clock 29,890 s for 5,000 PDFs
(5.98 s/PDF, 100% of budget), pass 1 complete on all 5,000 documents, and the
time-governed refinement pass covered the 757 highest-value worklist cases
before its deadline stop; the container exited cleanly on its own. The file
validates against scripts/validate_submission.py: 5,000 records, 0 missing
case ids. We state the lateness plainly rather than implying otherwise —
scoring eligibility is entirely at the organizers' discretion.