ToolPlanner: A Tool Augmented LLM for Multi Granularity Instructions with Path Planning and Feedback
It addresses the problem of aligning tool-augmented LLMs with real-world user instructions for researchers and practitioners, though it is incremental as it builds on existing tool-augmented LLM methods.
The paper tackles the gap between tool-augmented LLMs trained on overly detailed instructions and real-world user habits by introducing ToolPlanner, a two-stage reinforcement learning framework with path planning and feedback, which improves Match Rate, Pass Rate, and Win Rate by 26.8%, 20.2%, and 5.6% compared to the SOTA model.
Recently, tool-augmented LLMs have gained increasing attention. Given an instruction, tool-augmented LLMs can interact with various external tools in multiple rounds and provide a final answer. However, previous LLMs were trained on overly detailed instructions, which included API names or parameters, while real users would not explicitly mention these API details. This leads to a gap between trained LLMs and real-world scenarios. In addition, most works ignore whether the interaction process follows the instruction. To address these issues, we constructed a training dataset called MGToolBench, which contains statement and category-level instructions to better reflect real-world scenarios. In addition, we propose ToolPlanner, a two-stage reinforcement learning framework that utilizes path planning and two feedback mechanisms to enhance the LLM's task completion and instruction-following capabilities. Experimental results show that ToolPlanner significantly improves the Match Rate, Pass Rate and Win Rate by 26.8%, 20.2%, and 5.6% compared to the SOTA model. Human evaluation verifies that the multi-granularity instructions can better align with users' usage habits. Our data and code will be released upon acceptance.