arXiv cs.CL· Yihang Peng, Yi Jing, Dewen Liu, Weihan Meng, Ying Liu·· 2 天前AI 评分31
不是"是否"而是"何时"：从语境适切性诊断中文 LLM 的写作口头禅
Not X, but When: Diagnosing a Mandarin LLM Writing Tic through Contextual Appropriateness
AI 导读
研究以中文纠正构式"不是 X，而是 Y"为切口，用基于 Gricean 语用学的决策树将用法分为实质性、可选与缺陷三类。五款商用模型使用该构式的密度是人类参照的 2.5 至 9.2 倍；307 条标注用例中仅 1.30% 属缺陷，52.8% 属"正确但非必要"的可选用法，说明模型在适切性上的不足远多于正确性。
正文 · 原文
Abstract:Recurring expressions in model-written text are often treated as signatures of machine authorship, yet most of them serve legitimate functions. Whether such an expression is a tic depends on whether the model knows when it is called for. This judgment of appropriateness, central to communicative competence, remains largely unexamined for language models. We study it through the Mandarin corrective construction bushi X, ershi Y ("not X, but Y"). Drawing on Gricean pragmatics and corrective negation, we introduce a decision tree that classifies uses as substantive, optional, or defective by testing whether the context makes X a live alternative and whether rejecting X is informative. Five commercial models use the construction 2.5 to 9.2 times as densely as a human reference. Of 307 annotated uses, only 1.30% are defective while 52.8% are optional, that is, correct but unneeded, so models fall short in appropriateness far more often than in correctness. Several reward models favor the construction even over quality-matched variants, an LLM judge favors it even where the context does not license it, and both supervised fine-tuning and preference optimization tend to raise its frequency in open-weight models. We hope this work encourages the development of models attentive to the appropriateness of their expressions.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.22249 [cs.CL] |
| (or arXiv:2610.22249v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.22249 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yihang Peng [view email]
[v1]
Thu, 15 Oct 2026 17:22:11 UTC (1,855 KB)
来源：arXiv cs.CL · arxiv.org