arXiv：cs.CL· Olga Zamaraeva, Adri\'an Gude, Roi Santos-R\'ios, Carlos G\'omez-Rodr\'iguez·· 3 小时前AI 评分47
研究揭示 NLP 论文中「rather than」反义句式激增，GPT 生成论文尤甚
I would rather quit NLP than read another paper like this: The rise of antithesis in NLP papers
AI 导读
一项研究统计了 ACL 2019 论文、2026 年 ACL 风格 arXiv 论文及 GPT 生成论文中「rather than」句式的使用率，2026 年的使用率是 2019 年的七倍，GPT 论文中更高。
正文
Abstract:For better or worse, LLMs are by now used routinely for scientific writing.\footnote{This paper is no exception; we did use AI to assist with writing some of the sections (see Acknowledgments).} Many have noticed that recent models fill papers with unnecessary antithesis, stating over and over what the work does not do, in ways that do not contribute to its precision or quality of expression and annoy reviewers \emph{rather than impressing them}. We study the construction \emph{rather than} in ACL papers from 2019, ACL-style arXiv papers from 2026, and papers written by GPT models from the same titles and abstracts. Its rate in 2026 is seven times the 2019 rate, and higher still in the GPT papers. Two annotators, blind to the source, find almost no 2019 use \emph{annoying} and about one in ten 2026 uses; they seldom agree on which, yet about half of 2026 papers contain a use that annoys each of them. \emph{Annoying} uses present the rejected alternative less favorably than legitimate uses. Raters of preference data and open reward models favor the construction, and an instruction to be honest promotes it. We conjecture that it is a side effect of post-training on pairwise preferences, which credit a disavowal in a single response and cannot register its cost across a text.
| Comments: | 30 pages, 2 figures, 48 tables |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.10092 [cs.CL] |
| (or arXiv:2610.10092v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.10092 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Olga Zamaraeva [view email]
[v1]
Wed, 7 Oct 2026 13:50:50 UTC (95 KB)
来源：arXiv：cs.CL · arxiv.org