作者位次 #4/7ACL 2026Poster· Natural Language Processing
Learning from the Irrecoverable: Error-Localized Policy Optimization for Tool-Integrated LLM Reasoning
Qiao Liang · Yuke Zhu · Chao Ge · Lei Yang · Ying Shen · Bo Zheng (Taobao & · Sheng Guo
摘要
Tool-integrated reasoning (TIR) enables LLM agents to solve tasks through planning, tool use, and iterative revision, but outcome-only reinforcement learning in this setting suffers from sparse, delayed rewards and weak step-level credit assignment. In long-horizon TIR trajectories, an early irre...