Learning Situation-Conditioned Thinking Policies for Long-Term LLM Agents

作者：Hong Su · 单位：† † thanks: H. Su is with the School of Computer Science, Chengdu University of Information Technology, Chengdu, China. · 会议/期刊：arXiv preprint · 方向：cs.AI · 发布日期：2026-10-08