AWhat changed
Studied reasoning reinforcement learning without a supervised cold start in R1-Zero, then combined cold-start examples and staged training in R1 and distilled reasoning outputs to smaller models.
Scientific scope
Studied reasoning reinforcement learning without a supervised cold start in R1-Zero, then combined cold-start examples and staged training in R1 and distilled reasoning outputs to smaller models. R1-Zero's no-cold-start claim is not a claim about R1. GRPO is inherited. The paper reports language-mixing and general-capability limits; benchmark comparisons do not establish general superiority.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. R1-Zero's no-cold-start claim is not a claim about R1. GRPO is inherited. The paper reports language-mixing and general-capability limits; benchmark comparisons do not establish general superiority.
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.
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- DeepSeek-AI: R1 reasoning post-training and distillation — official research record
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