Search-R1：Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning

arXiv · UIUC, Google Cloud AI Research, UMass Am

将强化学习扩展到多轮检索与推理交织场景，通过检索token掩码与极简结果奖励让模型自主学会搜索与推理。