L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning

Open in new window