Tool-integrated Reinforcement Learning for Repo Deep Search
This addresses the problem of automating software issue localization for developers, representing an incremental improvement by integrating tools into LLM training.
The paper tackles the challenge of issue localization in software development, where natural language descriptions must be linked to faulty code through multi-step reasoning, by introducing ToolTrain, a training framework that enhances LLMs' ability to use retrieval tools, resulting in state-of-the-art performance with a 32B model surpassing Claude-3.7 on function-level localization and improved end-to-end issue resolution.
Issue localization, the process of identifying code locations that need modification to resolve software issues, is a critical yet challenging task in software development. The semantic gap between natural language issue descriptions and faulty code requires complex multi-hop reasoning through code dependencies. Existing LLM-based agents attempt to address this by integrating repository retrieval tools. However, this transforms issue localization into a demanding task we call Repo Deep Search, which requires the LLM to effectively utilize various repository retrieval tools throughout a multi-step reasoning and navigation process. To tackle this challenge, we present ToolTrain, a two-stage tool-integrated training framework combining rejection-sampled supervised fine-tuning and tool-integrated reinforcement learning to enhance LLMs' ability to use retrieval tools for issue localization. Experimental results show that ToolTrain-trained models achieve state-of-the-art performance, with our 32B model even surpassing Claude-3.7 on function-level localization. The results also show that improved localization performance translates to better end-to-end issue resolution performance. This further demonstrates that training for issue localization is a viable and effective strategy for improving automated software development.