用母语思维答题，模型跨语言答对率提升
When Language Shapes Thought: Cross-Lingual Transfer of Factual Knowledge in Question Answering
- 让模型用知识来源语言思考，而非默认英文
- 跨语言问答准确率平均提升5.2%，反超英语提示
- 适合多语言知识库、非英语主导场景的研究者
多语言大模型在跨语言信息获取中潜力巨大，但其事实知识的使用高度依赖输入语言。以往研究常采用英语提示与评估，假设英语推理具有普适优势。本文基于语言与思维理论，提出语言到思维（L2T）提示方法，使模型内部‘思考’语言与知识源语言一致。在三种语言和四种模型上验证，L2T表现持续优于英语提示，甚至逆转了英语优势。结果表明，跨语言知识迁移不应默认以英语为中心。代码已开源。
原文摘要 · Abstract (English)
Multilingual large language models (LLMs) offer promising opportunities for cross-lingual information access, yet their use of factual knowledge remains highly sensitive to the input language. Prior work has addressed this through English prompting and evaluation, assuming that English-based reasoning is universally beneficial. In this work, we challenge that assumption by exploring factual knowledge transfer from non-English to English through the lens of Language and Thought Theory. We introduce Language-to-Thought (L2T) prompting, which aligns the model's internal ''thinking'' language with the source of knowledge. Across three languages and four models, L2T consistently outperforms English-based reasoning, reversing the expected advantage of English prompts. Our code is available at https://github.com/GeomeunByeol/Language2Thought.
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。