用自然语言替代代码调用，让大模型更准地使用工具
Natural Language Tools: A Natural Language Approach to Tool Calling In Large Language Agents
- 用自然语言代替JSON格式调用工具，避免任务干扰
- 工具调用准确率提升18.4个百分点，输出波动降低70%
- 适合希望提升工具使用能力的开放模型用户
我们提出自然语言工具（NLT）框架，将大语言模型（LLMs）中程序化的JSON工具调用替换为自然语言输出。通过解耦工具选择与回复生成，NLT消除了任务干扰和格式约束，从而提升工具调用性能。在跨10个模型、6400次试验的客户服 务与心理健康领域评估中，NLT使工具调用准确率提高18.4个百分点，输出方差降低70%。开源模型获益最大，其表现超越部分旗舰闭源模型，对强化学习与监督微调阶段的训练均有启示。该优势在提示扰动下依然存在，并可扩展至无原生支持工具调用的模型。
原文摘要 · Abstract (English)
We present Natural Language Tools (NLT), a framework that replaces programmatic JSON tool calling in large language models (LLMs) with natural language outputs. By decoupling tool selection from response generation, NLT eliminates task interference and format constraints that degrade tool call performance. When evaluated across 10 models and 6,400 trials spanning customer service and mental health domains, NLT improves tool calling accuracy by 18.4 percentage points while reducing output variance by 70%. Open-weight models see the largest gains, surpassing flagship closed-weight alternatives, with implications for model training in both reinforcement learning and supervised fine-tuning stages. These improvements persist under prompt perturbations and extend tool-calling capabilities to models lacking native support.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。