arXiv cs.CL· Alison Chi, Sanghyeok Choi, Hannah Rohde, Emily Allaway·· 12 小时前AI 评分22
研究：CUES 线索如何驱动人类与 LLM 的典型性推理
Hey, guess what: CUES lead to inferences about typicality
AI 导读
研究构建了 EventTyp 与经人工验证的 EventTypGold 两个语料库，发现常识知识、话语强调与讽刺三条 CUES 线索可预测人类对事件典型性的推理。大规模人类与 LLM 实验显示，LLM 在多数实验设置下缺乏准确评估事件一般典型性所需的常识知识，因而难以遵循这些 CUES。
正文
Abstract:Throughout life, we accumulate knowledge of how typical certain events are. We use this knowledge in daily conversation, often interpreting meaning beyond what has been said explicitly. In this work, we take steps to better understand this meaning recovery, investigating both atypicality and typicality inferences. First, we construct two corpora, EventTyp and the human-validated EventTypGold. We conduct large-scale human and LLM experiments and find that for humans, three CUES predict inferences about typicality: Commonsense Knowledge, Utterance Emphasis, and Snark. However, in most experimental settings, we find that LLMs lack the prerequisite Commonsense Knowledge necessary to accurately assess an event's general typicality, and therefore struggle to follow these CUES.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.33239 [cs.CL] |
| (or arXiv:2610.33239v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.33239 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Alison Chi [view email]
[v1]
Fri, 23 Oct 2026 19:40:27 UTC (5,939 KB)
来源：arXiv cs.CL · arxiv.org