Tool Learning
Tool learning focuses on enabling large language models (LLMs) to interact with external tools to solve complex tasks beyond their inherent capabilities. Current research emphasizes improving the accuracy and efficiency of tool selection, invocation, and integration within the LLM's workflow, often employing techniques like generative models that embed tool knowledge directly into the model's parameters or methods that carefully balance tool-use improvements with maintaining general LLM performance. This field is significant because it bridges the gap between LLMs' reasoning abilities and the real world, paving the way for more robust and versatile AI agents capable of handling diverse, complex tasks across various domains.
Papers
From Exploration to Mastery: Enabling LLMs to Master Tools via Self-Driven Interactions
Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, Ji-Rong Wen
StepTool: A Step-grained Reinforcement Learning Framework for Tool Learning in LLMs
Yuanqing Yu, Zhefan Wang, Weizhi Ma, Zhicheng Guo, Jingtao Zhan, Shuai Wang, Chuhan Wu, Zhiqiang Guo, Min Zhang