发表机构
ETH Zürich; University of Zurich; UC San Diego（苏黎世联邦理工学院; 苏黎世大学; 加州大学圣地亚哥分校）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究用户信念表达形式对大语言模型的影响，引入基于语言维度的类型学，生成可控查询对，评估不同架构、规模和训练阶段的16个LLMs，发现响应行为差异及能显著说服模型的特定信念表达，揭示语言框架对模型情境整合的影响。
AI 中文摘要
用户经常向大语言模型（LLMs）表达他们的信念。在某些情况下，LLM应该接受这些情境信念为真。在其他情况下，它们应坚持其先验知识。用户的信念表达（EoBs）在语言形式上多种多样，使用预设、证据和确定性标记或不同语调等，每种形式对LLMs的说服力可能不同。我们引入一种类型学来系统评估不同EoBs如何影响模型遵循情境还是先验知识。该类型学基于四个语言维度：形式、证据性、认知立场和语调，涵盖17种细粒度类型。通过将这些EoBs与世界知识事实配对，生成可控的EoB查询对以分离语言变化的影响。使用此基准评估16个在架构、规模和训练阶段不同的LLMs。我们识别出这些方面响应行为的有意义差异，还确定了比其他EoBs更能显著说服LM的特定EoBs。我们的工作揭示了语言框架影响LLM情境整合的系统模式，对提示工程和模型鲁棒性有影响。
英文摘要
Users frequently express their beliefs to large language models (LLMs). In some situations, the LLM should accept these contextual beliefs as true. In others, they should stick to their prior knowledge. Notably, users' expressions of belief (EoBs) can take linguistically diverse forms - using presuppositions, evidential and certainty markers, or varied tones - each of which may have a different persuasiveness over the LLMs. We introduce a typology to systematically evaluate how different EoBs affect whether models follow context versus prior knowledge. The typology is grounded in four linguistically motivated dimensions: form, evidentiality, epistemic stance, and tone, spanning 17 fine-grained types. By pairing these EoBs with world knowledge facts, we generate controlled EoB-query pairs that isolate the effect of linguistic variation. Using this benchmark, we evaluate 16 LLMs that differ in architecture (Llama3, Qwen3, Gemma3), scale (1B-30B parameters), and training stages (base vs instruct). We identify meaningful variations in response behavior across these axes, e.g., that bigger models and instruction models tend to be less context-following than smaller models and base models. We further identify specific EoBs that statistically significantly persuade LMs more consistently than others. Our work reveals systematic patterns in how linguistic framing affects LLM context integration, with implications for prompt engineering and model robustness.
CommentsPublished at ACL 2026
DOI：10.18653/v1/2026.acl-long.142
