paper-with-me

홈 › Papers

StepTool: A Step-grained Reinforcement Learning Framework for Tool Learning in LLMs

2024-10-10 · Yuanqing Yu, Zhefan Wang, Weizhi Ma, Zhicheng Guo, Jingtao Zhan, Shuai Wang, Chuhan Wu, Zhiqiang Guo, Min Zhang

Despite having powerful reasoning and inference capabilities, Large Language Models (LLMs) still need external tools to acquire real-time information retrieval or domain-specific expertise to solve complex tasks, which is referred to as tool learning. Existing tool learning methods primarily rely on tuning with expert trajectories, focusing on token-sequence learning from a linguistic perspective. However, there are several challenges: 1) imitating static trajectories limits their ability to generalize to new tasks. 2) even expert trajectories can be suboptimal, and better solution paths may exist. In this work, we introduce StepTool, a novel step-grained reinforcement learning framework to improve tool learning in LLMs. It consists of two components: Step-grained Reward Shaping, which assigns rewards at each tool interaction based on tool invocation success and its contribution to the task, and Step-grained Optimization, which uses policy gradient methods to optimize the model in a multi-step manner. Experimental results demonstrate that StepTool significantly outperforms existing methods in multi-step, tool-based tasks, providing a robust solution for complex task environments. Codes are available at https://github.com/yuyq18/StepTool.

📄 PDF Abstract BibTeX arXiv:2410.07745

Code (1)

yuyq18/steptool 공식 구현 pytorch

Tasks

Information RetrievalPolicy Gradient Methods

Similar Papers 제목 키워드 기반

MatchTIR: Fine-Grained Supervision for Tool-Integrated Reasoning via Bipartite Matching

2026-01-15 · Changle Qu, Sunhao Dai, Hengyi Cai, Jun Xu 외 arxiv

Tool-Integrated Reasoning (TIR) empowers large language models (LLMs) to tackle complex tasks by interleaving reasoning steps with external tool interactions. However, existing reinforcement learning methods typically re…

Reinforcement Learning

Every Step Counts: Step-Level Credit Assignment for Tool-Integrated Text-to-SQL

2026-05-06 · Yaxun Dai, Baolin Sun, Junying Wang, Pengfei Wang 외 arxiv

Tool-integrated Text-to-SQL parsing has emerged as a promising paradigm, framing SQL generation as a sequential decision-making process interleaved with tool execution. However, existing reinforcement learning approaches…

Reinforcement Learning

Ophiuchus: Incentivizing Tool-augmented "Think with Images" for Joint Medical Segmentation, Understanding and Reasoning

2025-12-16 · Yankai Jiang, Yujie Zhang, Peng Zhang, Wenjie Li 외 arxiv

Recent medical MLLMs have made significant progress in generating step-by-step textual reasoning chains. However, they still struggle with complex clinical tasks that necessitate dynamic and iterative focusing on fine-gr…

Reinforcement Learning

CM2: Reinforcement Learning with Checklist Rewards for Multi-Turn and Multi-Step Agentic Tool Use

2026-02-12 · Zhen Zhang, Kaiqiang Song, Xun Wang, Yebowen Hu 외 arxiv

AI agents are increasingly used to solve real-world tasks by reasoning over multi-turn user interactions and invoking external tools. However, applying reinforcement learning to such settings remains difficult: realistic…

Reinforcement Learning

TIER: Trajectory-Invariant Execution Rewards for Multi-Step Tool Composition

2026-05-16 · Anay Kulkarni, ChiaEn Lu, Dheeraj Mekala, Jayanth Srinivasa 외 arxiv

Tool use enables large language models to solve complex tasks through sequences of API calls, yet existing reinforcement learning approaches fail to scale to multi-step composition settings. Outcome-based rewards provide…

Reinforcement Learning