ReMA: Learning to Meta-think for LLMs with Multi-Agent Reinforcement Learning

作者：Ziyu Wan, Yunxiang Li, Xiaoyu Wen, Yan Song, Hanjing Wang, Linyi Yang, Mark Schmidt, Jun Wang, Weinan Zhang, Shuyue Hu, Ying Wen · 单位：Shanghai Jiao Tong University, Shanghai Artificial Intelligence Laboratory, University of British Columbia, University College London · 会议/期刊：arXiv preprint · 方向：智能体架构 · 发布日期：2025-03-12