arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.12435cs.CL
收藏

语言最小对立体引发大语言模型中的语言相似性

Linguistic Minimal Pairs Elicit Linguistic Similarity in Large Language Models

Université Paris Cité(巴黎西岱大学);Sorbonne Université(索邦大学);CUHK(香港中文大学);HKUST(香港科技大学);Cohere;Brain and Mind Institute, CUHK(香港中文大学脑与心智研究所)

Xinyu Zhou, Delong Chen, Samuel Cahyawijaya, Xufeng Duan, Zhenguang G. Cai

AI总结:

本研究利用语言最小对立体探究大语言模型的内部语言表征,通过大规模实验从四方面揭示语言相似性特性,阐明了LLM与语言学理论的关系。

摘要:

我们提出了一种新颖的分析方法,利用语言最小对立体来探究大语言模型(LLM)的内部语言表征。通过测量最小对立体之间LLM激活差异的相似性,我们对LLM捕获的语言知识进行了量化并获得了深入见解。我们的大规模实验涵盖100多个LLM和三种语言的15万个最小对立体,从四个关键方面揭示了语言相似性的特性:LLM之间的一致性、与理论分类的关系、对语义上下文的依赖性以及相关现象的跨语言对齐。我们的研究结果表明:1)语言相似性受训练数据接触量的显著影响,导致资源更丰富的语言中LLM间的一致性更高。2)语言相似性与细粒度的理论语言类别高度一致,但与更宽泛的类别一致性较弱。3)语言相似性与语义相似性呈弱相关,体现出其上下文依赖的特性。4)LLM在对相关语言现象的理解上表现出有限的跨语言对齐。这项工作证明了最小对立体作为洞察LLM中语言神经表征窗口的潜力,阐明了LLM与语言学理论之间的关系。代码和数据可在https://github.com/ChenDelong1999/Linguistic-Similarity获取。

英文摘要:

We introduce a novel analysis that leverages linguistic minimal pairs to probe the internal linguistic representations of Large Language Models (LLMs). By measuring the similarity between LLM activation differences across minimal pairs, we quantify the and gain insight into the linguistic knowledge captured by LLMs. Our large-scale experiments, spanning 100+ LLMs and 150k minimal pairs in three languages, reveal properties of linguistic similarity from four key aspects: consistency across LLMs, relation to theoretical categorizations, dependency to semantic context, and cross-lingual alignment of relevant phenomena. Our findings suggest that 1) linguistic similarity is significantly influenced by training data exposure, leading to higher cross-LLM agreement in higher-resource languages. 2) Linguistic similarity strongly aligns with fine-grained theoretical linguistic categories but weakly with broader ones. 3) Linguistic similarity shows a weak correlation with semantic similarity, showing its context-dependent nature. 4) LLMs exhibit limited cross-lingual alignment in their understanding of relevant linguistic phenomena. This work demonstrates the potential of minimal pairs as a window into the neural representations of language in LLMs, shedding light on the relationship between LLMs and linguistic theory. Codes and data are available at https://github.com/ChenDelong1999/Linguistic-Similarity

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。

↑