Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?
Yue et al., 2025, Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? (arXiv preprint arXiv:2504.13837)
Subscriber deck — free during early access
1 cards in “1. Introduction” · whole deck
1. Introduction
- What widely held belief about RL with verifiable rewards (RLVR) do Yue et al. test?That, like RL in Go or Atari, RLVR lets LLMs discover genuinely new reasoning strategies beyond their base models. The alternative they test: it only makes the model better at finding reasoning paths the base model could already produce.
- Yue et al., 2025, Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? (arXiv preprint arXiv:2504.13837), 1. Introduction
Source
- Yang Yue, Zhiqi Chen, Rui Lu, Andrew Zhao, Zhaokai Wang, Yang Yue, Shiji Song and Gao Huang. 2025. Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?. In arXiv preprint arXiv:2504.13837. https://arxiv.org/abs/2504.13837