AWhat changed
Compared step-level, outcome-based and learned-reward feedback for GSM8K reasoning, separating final-answer accuracy from correctness of the reasoning trace.
Scientific scope
Compared step-level, outcome-based and learned-reward feedback for GSM8K reasoning, separating final-answer accuracy from correctness of the reasoning trace. Outcome supervision reached similar final-answer errors with fewer labels in this setup; process-based feedback improved reasoning correctness. Neither this result nor the later MATH study establishes a universal winner across tasks.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Outcome supervision reached similar final-answer errors with fewer labels in this setup; process-based feedback improved reasoning correctness. Neither this result nor the later MATH study establishes a universal winner across tasks.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction