为阿拉伯语大模型构建工具调用能力，验证了本地数据与指令微调的有效性。
Tool Calling for Arabic LLMs: Data Strategies and Instruction Tuning
- 使用阿拉伯语数据集进行指令微调，提升模型对工具调用的理解能力。
- 在阿拉伯语模型上微调特定高优先级工具，性能提升显著。
- 研究结果为多语言大模型工具调用提供了可复用的训练策略。
工具调用是大型语言模型（LLMs）与外部系统交互的关键能力，极大拓展了其应用范围。然而，现有研究和资源主要集中于英语，导致阿拉伯语等语言的工具调用能力发展滞后。本文围绕三个核心问题展开：(1) 是否必须依赖阿拉伯语原生工具调用数据，而非跨语言迁移；(2) 通用指令微调对工具调用性能的影响；(3) 针对特定高优先级工具进行微调的价值。为此，我们基于开源阿拉伯语LLM的基线与后训练版本进行了广泛实验。为支持研究，我们首次将两个开源工具调用数据集翻译并适配为阿拉伯语。研究结果揭示了在阿拉伯语中构建稳健工具增强型智能体的最佳策略。
原文摘要 · Abstract (English)
Tool calling is a critical capability that allows Large Language Models (LLMs) to interact with external systems, significantly expanding their utility. However, research and resources for tool calling are predominantly English-centric, leaving a gap in our understanding of how to enable this functionality for other languages, such as Arabic. This paper investigates three key research questions: (1) the necessity of in-language (Arabic) tool-calling data versus relying on cross-lingual transfer, (2) the effect of general-purpose instruction tuning on tool-calling performance, and (3) the value of fine-tuning on specific, high-priority tools. To address these questions, we conduct extensive experiments using base and post-trained variants of an open-weight Arabic LLM. To enable this study, we bridge the resource gap by translating and adapting two open-source tool-calling datasets into Arabic. Our findings provide crucial insights into the optimal strategies for developing robust tool-augmented agents for Arabic.
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。