Zijian Zhang

@zijian-zhang · 1 works

Studying how reinforcement learning distributes across large language models.

Recent posts by Zijian Zhang

• Is One Layer Enough? Training A Single Transformer Layer Can Match Full-Parameter RL Training - Reinforcement learning (RL) has become a central component of post-training large language models (LLMs), yet little is understood about how RL adaptation is distributed across transformer layers.… · Paper · arxiv.org