imadreamerboy/repro-reinforcement-learning-for-rl-algorithms

RL4RLA: bounded source/table audit

This static candidate covers OpenReview Oj2I1xdKpv, RL4RLA: Teaching ML to Discover Randomized Linear Algebra Algorithms Through Curriculum Design and Graph-Based Search.

Evidence limit: this package contains a released-source and paper-table audit plus one runner dry-run; it is not an independent reproduction and does not validate the reported numerical results.

The exact active judge claims were extracted at challenge revision 7b5b56aebf3abe590eab9f2c241a796125cab928. The judge loads both feeds and applies the anchored map last, so the six anchored claims replace the two default claims for this OpenReview ID. The default records are retained only as immutable source inputs. This package contains only the target's extracted claim records, direct public links, and newly authored audit text. It intentionally excludes the paper PDF/source archive, author repository, data, binaries, logs, and any local checkout.

Run the offline release check with:

python3 scripts/verify_release.py

See the claim matrix and evidence boundaries.