Feedback-Driven Tool-Use Improvements in Large Language Models via Automated Build Environments
This addresses a bottleneck in improving LLM interaction with environments, offering a novel method for training but is incremental in advancing tool-use capabilities.
The paper tackles the lack of efficient reinforcement learning frameworks for tool use in large language models by proposing an automated environment construction pipeline and a verifiable reward mechanism, which significantly enhances tool-use performance without degrading general capabilities across varying model scales.
Effective tool use is essential for large language models (LLMs) to interact meaningfully with their environment. However, progress is limited by the lack of efficient reinforcement learning (RL) frameworks specifically designed for tool use, due to challenges in constructing stable training environments and designing verifiable reward mechanisms. To address this, we propose an automated environment construction pipeline, incorporating scenario decomposition, document generation, function integration, complexity scaling, and localized deployment. This enables the creation of high-quality training environments that provide detailed and measurable feedback without relying on external tools. Additionally, we introduce a verifiable reward mechanism that evaluates both the precision of tool use and the completeness of task execution. When combined with trajectory data collected from the constructed environments, this mechanism integrates seamlessly with standard RL algorithms to facilitate feedback-driven model training. Experiments on LLMs of varying scales demonstrate that our approach significantly enhances the models' tool-use performance without degrading their general capabilities, regardless of inference modes or training algorithms. Our analysis suggests that these gains result from improved context understanding and reasoning, driven by updates to the lower-layer MLP parameters in models.