SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning

作者：Bo Liu, Leon Guertler, Simon Yu, Zichen Liu, Penghui Qi, Daniel Balcells, Mickel Liu, Cheston Tan, Weiyan Shi, Min Lin, Wee Sun Lee, Natasha Jaques · 单位：National University of Singapore Northeastern University Sea AI Lab, Centre for Frontier AI Research (CFAR), A STAR Plastic Labs University of Washington · 会议/期刊：arXiv preprint · 方向：奖励驱动进化 · 发布日期：2025-06-30