arXiv cs.CL· Yuchen Li, Lisa Liu, Runqi Zou, Shuxuan Kuang, Qingquan Zhang, Julian Togelius·· 11 小时前AI 评分22
语言作为特征、礼貌作为策略：LLM 社交推理智能体语言效果的受控测试
Language as Trait, Politeness as Strategy: Controlled Tests of Linguistic Effects in LLM Social-Deduction Agents
AI 导读
研究用 LURK Werewolf 框架对 LLM 智能体做受控测试，只改写话语的语言实现而保留言语行为内容，检验语言复杂度与礼貌策略是否影响其他智能体的判断与行动。
正文 · 原文
Abstract:Large language model (LLM) agents coordinate through conversation, and the same message can be phrased in simpler or more complex language, bluntly or politely. We test whether such differences change how other agents judge and act, using the LURK Werewolf framework, which separates each utterance's speech-act predicate from its natural-language realization. We rewrite only the realization, check that the game content is preserved, and replay each game day from saved checkpoints against a re-sampled control. We vary linguistic range consistently across a speaker's messages and politeness strategy per utterance. Across 30 games and 4,335 replays, neither manipulation changes the trust other agents place in the speaker, and no effect on votes or night actions survives correction for multiple comparisons; the largest uncorrected effect is that shifting range away from a neutral paraphrase changes about 4 percentage points more of the other players' votes. In natural play, 88% of face-threatening speech is bald on-record regardless of role.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.09768 [cs.CL] |
| (or arXiv:2611.09768v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.09768 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yuchen Li [view email]
[v1]
Fri, 6 Nov 2026 17:31:01 UTC (70 KB)
来源：arXiv cs.CL · arxiv.org