Fetching the paper…

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