arXiv：cs.CL· Michael A. Lepori, Jennifer Hu, Ishita Dasgupta, Roma Patel, Thomas Serre, Ellie Pavlick·· 16 小时前
语言模型能否判断事件可能性？研究揭示模态差异向量
Is This Just Fantasy? Language Model Representations Reflect Human Judgments of Event Plausibility
AI 导读
研究在多种语言模型内部识别出可区分句子模态类别（可能、不可能、完全无意义等）的线性表示，称为模态差异向量，其可靠性高于此前报告的水平。这些向量随训练步数、层数和参数量的增加按一致顺序出现，且可用于建模人类对模态类别的细粒度判断行为。
正文
Abstract:Language models (LMs) are used for a diverse range of tasks, from question answering to writing fantastical stories. In order to reliably accomplish these tasks, LMs must be able to discern the modal category of a sentence (i.e., whether it describes something that is possible, impossible, completely nonsensical, etc.). However, recent studies have called into question the ability of LMs to categorize sentences according to modality (Michaelov et al., 2025; Kauf et al., 2023). In this work, we identify linear representations that discriminate between modal categories within a variety of LMs, or modal difference vectors. Analysis of modal difference vectors reveals that LMs have access to more reliable modal categorization judgments than previously reported. Furthermore, we find that modal difference vectors emerge in a consistent order as models become more competent (i.e., through training steps, layers, and parameter count). Notably, we find that modal difference vectors identified within LM activations can be used to model fine-grained human categorization behavior. This potentially provides a novel view into how human participants distinguish between modal categories, which we explore by correlating projections along modal difference vectors with human participants' ratings of interpretable features. In summary, we derive new insights into LM modal categorization using techniques from mechanistic interpretability, with the potential to inform our understanding of modal categorization in humans.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2507.12553 [cs.CL] |
| (or arXiv:2507.12553v4 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2507.12553 arXiv-issued DOI via DataCite |
Submission history
From: Michael Lepori Jr. [view email]
[v1]
Wed, 16 Jul 2025 18:04:26 UTC (659 KB)
[v2]
Wed, 25 Feb 2026 23:03:52 UTC (675 KB)
[v3]
Mon, 27 Apr 2026 19:53:17 UTC (681 KB)
[v4]
Tue, 13 Oct 2026 21:05:35 UTC (681 KB)
来源：arXiv：cs.CL · arxiv.org