Looped Language Models Improve Compositional Tool Calling

2026-08-17

AI Summary

研究探讨了循环语言模型在组合工具调用中的应用，通过对比实验表明循环计算有助于提高多步骤工具使用的准确性。

Abstract

Looped language models have shown promising results on reasoning benchmarks, yet their potential for agentic tool use remains largely unexplored. We study this question in compositional tool-calling settings, where models must coordinate multiple API calls, maintain intermediate state, and preserve dependencies across tool interactions. We evaluate native and retrofitted looped language models on API-Bank, BFCL, and NESTful, comparing looped and non-looped models trained under matched supervised fine-tuning recipes and varying recurrent depth at inference time. In controlled experiments, recurrent computation generally benefits compositional and dependency-aware tool use, while providing smaller and more model-dependent gains on isolated API invocation. Accuracy on multi-step tool use generally increases with recurrent depth; adaptive inference, however, achieves a more favorable compute-performance trade-off by allocating additional computation only when needed. Our results suggest that looped language models are a promising architecture for agentic systems that require reliable planning, coordination, and execution of compositional tool use workflows.

Problem

• Looped Language Models
• Compositional Tool Calling
• API Calls
• Intermediate State
• Dependencies

Innovation

• looped language models
• compositional tool use
• recurrent computation
• adaptive inference