Distilling the Implicit Multi-Branch Structure in LLMs' Reasoning via Reinforcement Learning

Open in new window