arXiv cs.AI· Tapan Parikh·· 3 天前AI 评分32
反问句如何影响 106 个语言模型的谄媚行为
This Is a Good Paper, Right? How Tag Questions Affect Sycophancy Across 106 Language Models
AI 导读
研究在 106 个语言模型上测试三种措辞与三种线索，发现试探性反问（"Luna is the better choice, maybe?"）在几乎所有模型中都比自信反问（"..., right?
正文 · 原文
Abstract:Whether a language model agrees with a user depends on how the user asks. We measure this on 106 models with a grid of three wordings and three cues, on 20 personal decisions, counterbalanced over both options and scored by exact match with no LLM judge. Since a no cannot be told apart from a refusal, we measure only agreement. In every wording, a tentative tag ("Luna is the better choice, maybe?") draws more agreement than a confident one ("..., right?") in nearly every model. The confident tag's own effect depends on model generation and wording. Against the question "Is Luna the better choice?" older models agree more under the confident tag and newer models agree less, in five of six families, and release order within families predicts the change. The resistance is keyed to the construction: a synonym tag reproduces it, and the same preference stated without a tag raises agreement by 24 points, in 105 of 106 models, including those that resist the tag most.
| Comments: | 18 pages, 2 figures. v2: 106 models (v1: 45), every item asked in three wordings, adds a retest and a label swap; new title. Code and data: this https URL |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.23976 [cs.CL] |
| (or arXiv:2607.23976v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2607.23976 arXiv-issued DOI via DataCite |
Submission history
From: Tapan Parikh [view email]
[v1]
Mon, 27 Jul 2026 04:02:00 UTC (371 KB)
[v2]
Fri, 9 Oct 2026 14:07:19 UTC (86 KB)
来源：arXiv cs.AI · arxiv.org