测试大模型能否像人一样用常识判断句子歧义，发现它们表现不佳。
Plausibility as Commonsense Reasoning: Humans Succeed, Large Language Models Do not
- 用土耳其语相对从句歧义句测试模型对常识的依赖。
- 人类能正确利用事件合理性判断句义，大模型则效果差或方向相反。
- 适合研究语言理解机制与人类认知差异的研究者参考。
大型语言模型在诸多语言任务中表现优异，但其在歧义消解时是否以类人方式、结合世界知识与句法结构进行判断仍不明确。本研究聚焦土耳其语中前置定语从句的歧义现象，同一表面结构可支持高附着（HA）或低附着（LA）两种解析。我们构建了句法配置固定、两种解析均语用合理的歧义句，并通过事件合理性的梯度变化，使高附着或低附着更可信。实验采用独立评分验证对比差异。在快速强制选择的语义理解实验中，人类表现出显著且方向正确的合理性效应。随后，在平行偏好设置下，评估土耳其语及多语言大模型，通过平均每词对数概率比较匹配的HA/LA延续。结果显示，各模型的合理性驱动偏差微弱、不稳定甚至反向。表明当前模型在处理此类语义歧义时，无法像人类一样可靠地利用合理性信息引导附着判断，凸显土耳其语从句附着作为跨语言诊断工具的价值。
原文摘要 · Abstract (English)
Large language models achieve strong performance on many language tasks, yet it remains unclear whether they integrate world knowledge with syntactic structure in a human-like, structure-sensitive way during ambiguity resolution. We test this question in Turkish prenominal relative-clause attachment ambiguities, where the same surface string permits high attachment (HA) or low attachment (LA). We construct ambiguous items that keep the syntactic configuration fixed and ensure both parses remain pragmatically possible, while graded event plausibility selectively favors High Attachment vs.\ Low Attachment. The contrasts are validated with independent norming ratings. In a speeded forced-choice comprehension experiment, humans show a large, correctly directed plausibility effect. We then evaluate Turkish and multilingual LLMs in a parallel preference-based setup that compares matched HA/LA continuations via mean per-token log-probability. Across models, plausibility-driven shifts are weak, unstable, or reversed. The results suggest that, in the tested models, plausibility information does not guide attachment preferences as reliably as it does in human judgments, and they highlight Turkish RC attachment as a useful cross-linguistic diagnostic beyond broad benchmarks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。