2025-04-23 13:51:18.AIbase.17.5k
Academic Fraud Busting! Research from Tsinghua and SJTU Upends Understanding: Reinforcement Learning May Hinder Large Model Reasoning
A recent paper jointly published by Tsinghua University and Shanghai Jiao Tong University challenges the widely held belief that pure reinforcement learning (RL) enhances large model reasoning capabilities. The research found that models incorporating reinforcement learning performed worse than their original counterparts in certain tasks.
