arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2505.03368cs.LG
收藏

大型语言模型的地理空间机制可解释性

Geospatial Mechanistic Interpretability of Large Language Models

University of Leicester, UK(利兹大学);University of Udine, Italy(乌迪内大学)

Stef De Sabbata, Stefano Mizzaro, Kevin Roitero

AI总结:

本文提出地理空间机制可解释性框架,结合探测、稀疏自编码器与空间自相关分析,揭示 LLM 地名特征中的地理位置结构。

摘要:

大型语言模型(LLM)已在各种自然语言处理任务中展现出前所未有的能力。它们处理并生成可行文本和代码的能力使其在许多领域无处不在,而将其作为知识库和“推理”工具部署仍是一个持续研究的领域。在地理学中,越来越多的文献一直专注于评估 LLM 的地理知识及其执行空间推理的能力。然而,关于这些模型的内部运作,尤其是它们如何处理地理信息,人们仍然知之甚少。在本章中,我们建立了一个研究地理空间机制可解释性的新颖框架——使用空间分析来逆向工程 LLM 如何处理地理信息。我们的目标是增进对这些复杂模型在处理地理信息时生成的内部表征的理解——如果这种说法不是不恰当的拟人化,人们或许可以称之为“LLM 如何思考地理信息”。我们首先概述探测(probing)在揭示 LLM 内部结构方面的应用。随后介绍机制可解释性领域,讨论叠加假设以及稀疏自编码器在将 LLM 的多语义内部表征解耦为更具可解释性的单语义特征方面的作用。在实验中,我们使用空间自相关来说明针对地名获得的特征如何呈现与其地理位置相关的空间模式,从而能够在地理空间上得到解释,为理解这些模型如何处理地理信息提供洞见。最后,我们讨论该框架如何帮助塑造地理学中基础模型的研究与使用。

英文摘要:

Large Language Models (LLMs) have demonstrated unprecedented capabilities across various natural language processing tasks. Their ability to process and generate viable text and code has made them ubiquitous in many fields, while their deployment as knowledge bases and "reasoning" tools remains an area of ongoing research. In geography, a growing body of literature has been focusing on evaluating LLMs' geographical knowledge and their ability to perform spatial reasoning. However, very little is still known about the internal functioning of these models, especially about how they process geographical information. In this chapter, we establish a novel framework for the study of geospatial mechanistic interpretability - using spatial analysis to reverse engineer how LLMs handle geographical information. Our aim is to advance our understanding of the internal representations that these complex models generate while processing geographical information - what one might call "how LLMs think about geographic information" if such phrasing was not an undue anthropomorphism. We first outline the use of probing in revealing internal structures within LLMs. We then introduce the field of mechanistic interpretability, discussing the superposition hypothesis and the role of sparse autoencoders in disentangling polysemantic internal representations of LLMs into more interpretable, monosemantic features. In our experiments, we use spatial autocorrelation to show how features obtained for placenames display spatial patterns related to their geographic location and can thus be interpreted geospatially, providing insights into how these models process geographical information. We conclude by discussing how our framework can help shape the study and use of foundation models in geography.

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。

↑