The Thinking Leakage in Reinforcement Learning for LLMs

7 points | by ilreb 9 hours ago