examples: OpenAI RFT python-grader template (IPT isomorphic reward) - #5
Merged
Conversation
Self-contained grade(sample,item)->float for the OpenAI RFT grader sandbox: rewards the ISOMORPHIC re-grade (fresh inputs recomputed from a trusted reference) so an RL policy cannot farm reward by memorizing the provided tests. stdlib only, deterministic, no network. Self-test asserts genuine=1.0 > memorizer(shortcut)=0.0. Honest scope: needs trusted reference+generator; IPT is public prior art (arXiv:2604.15149); GRPO prevention number labeled as our reproduced toy. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Self-contained IPT reward grader for the OpenAI RFT sandbox. Rewards the isomorphic re-grade so memorizing provided tests earns 0. stdlib-only, deterministic, self-test (genuine 1.0 > memorizer 0.0). Honesty-scoped.