On Predictability of Reinforcement Learning Dynamics for Large Language Models

Re-assign community