arXiv 2504.101912025-04-14cs.CLcs.AI

大型语言模型中本地化文化知识的保守性与可控性

Localized Cultural Knowledge is Conserved and Controllable in Large Language Models

Princeton University（普林斯顿大学）；EPFL（洛桑联邦理工学院）；Northeastern University（东北大学）；University of Chicago（芝加哥大学）；Santa Fe Institute（圣塔菲研究所）

Veniamin Veselovsky, Berke Argin, Benedikt Stroebl, Chris Wendler, Robert West, James Evans, Thomas L. Griffiths, Arvind Narayanan

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本研究针对LLMs多语言交互中默认英语中心的问题，通过显式文化背景提示缩小显隐本地化差距，识别跨非英语语言的文化定制向量，保留回应多样性并减少刻板印象，为文化定制及模型功能扩展提供新方向。

摘要：

正如人类在说新语言时会表现出受母语影响的语言模式一样，大型语言模型（LLMs）即使在生成其他语言的内容时，也常常默认以英语为中心的回应。尽管如此，我们观察到模型中仍保留着本地文化信息，并且可以很容易地激活这些信息以进行文化定制。我们首先证明，在提示词中明确提供文化背景能显著提高模型生成文化本地化回应的能力。我们将模型在有和没有明确文化背景时的性能差异称为显式-隐式本地化差距，这表明虽然LLMs中存在文化知识，但如果不明确提供文化背景，这些知识在多语言交互中可能不会自然显现。然而，尽管显式提示有好处，生成的答案多样性会降低，且容易出现刻板印象。其次，我们识别出一个显式文化定制向量，该向量在我们探索的所有非英语语言中都保持一致，它能使LLMs从合成的英语文化世界模型转向每种非英语文化世界。经过引导的回应保留了隐式提示的多样性，并减少了刻板印象，从而显著提高了定制潜力。我们讨论了显式文化定制对理解LLMs中替代文化世界模型的保守性的意义，以及它们在翻译、文化定制方面的可控效用，还有通过软控制使显式知识隐式化以扩展LLM功能和吸引力的可能性。

英文摘要：

Just as humans display language patterns influenced by their native tongue when speaking new languages, LLMs often default to English-centric responses even when generating in other languages. Nevertheless, we observe that local cultural information persists within the models and can be readily activated for cultural customization. We first demonstrate that explicitly providing cultural context in prompts significantly improves the models' ability to generate culturally localized responses. We term the disparity in model performance with versus without explicit cultural context the explicit-implicit localization gap, indicating that while cultural knowledge exists within LLMs, it may not naturally surface in multilingual interactions if cultural context is not explicitly provided. Despite the explicit prompting benefit, however, the answers reduce in diversity and tend toward stereotypes. Second, we identify an explicit cultural customization vector, conserved across all non-English languages we explore, which enables LLMs to be steered from the synthetic English cultural world-model toward each non-English cultural world. Steered responses retain the diversity of implicit prompting and reduce stereotypes to dramatically improve the potential for customization. We discuss the implications of explicit cultural customization for understanding the conservation of alternative cultural world models within LLMs, and their controllable utility for translation, cultural customization, and the possibility of making the explicit implicit through soft control for expanded LLM function and appeal.

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。