语言优先于内容：追踪多语言大型语言模型中的文化理解
Language over Content: Tracing Cultural Understanding in Multilingual Large Language Models
AI总结：
本研究通过测量激活路径重叠度，探究多语言大型语言模型的内部文化理解机制，发现语言对内部表示的影响强于文化，且语言相似性未必带来文化表示一致性。
摘要：
大型语言模型（LLMs）正越来越多地应用于不同文化背景中，因此准确的文化理解至关重要。此前的评估大多聚焦于输出层面的性能，掩盖了导致响应差异的因素，而使用电路分析的研究则覆盖语言较少且极少关注文化。本研究通过测量两种条件下回答语义等价问题时的激活路径重叠度，追踪LLMs的内部文化理解机制：固定问题语言而改变目标国家，以及固定国家而改变问题语言。我们还使用同语言国家对来区分语言与文化层面。结果显示，同语言跨国家问题的内部路径重叠度高于跨语言同国家问题，表明存在强烈的语言特定模式。值得注意的是，韩国-朝鲜对重叠度低且变异性高，说明语言相似性不能保证内部表示一致。
英文摘要：
Large language models (LLMs) are increasingly used across diverse cultural contexts, making accurate cultural understanding essential. Prior evaluations have mostly focused on output-level performance, obscuring the factors that drive differences in responses, while studies using circuit analysis have covered few languages and rarely focused on culture. In this work, we trace LLMs' internal cultural understanding mechanisms by measuring activation path overlaps when answering semantically equivalent questions under two conditions: varying the target country while fixing the question language, and varying the question language while fixing the country. We also use same-language country pairs to disentangle language from cultural aspects. Results show that internal paths overlap more for same-language, cross-country questions than for cross-language, same-country questions, indicating strong language-specific patterns. Notably, the South Korea-North Korea pair exhibits low overlap and high variability, showing that linguistic similarity does not guarantee aligned internal representation.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
