arXiv cs.CL· Wiem Ben Rim, Yao Lu, Pontus Stenetorp, Emine Yilmaz·· 12 小时前AI 评分38
俄语、葡萄牙语、德语提问同一争议话题：LLM 观点多样性跨语言画像研究
A Russian, a Brazilian, and a German Walk into a Bar, Ask the Same Question-- And Get Three Different Answers: Profiling LLM Perspective Diversity on Contested Topics
AI 导读
研究者构建了覆盖 18 种语言、经人工核验的多语言争议话题数据集 PerspectivesQA，并提出 ROOTS 框架刻画语言、观点与表述框架的多样性。对 8 个开源与闭源模型的评估显示，LLM 在英语中倾向模糊表态，在低资源语言中却给出明确立场；大模型跨语言世界观更一致，小模型更易被左右，且所有模型都对输入表层微小变化高度敏感。
正文
Abstract:Large Language Models(LLMs) have shown strong performance on tasks with verifiable answers, but many real-world queries involve subjectivity, disagreement, and conflicting evidence. Most existing benchmarks for such queries remain English-centric, overlooking the vast diversity of perspectives held by users worldwide. To systematically study how LLM perspectives vary across languages, we curate PerspectivesQA, a manually verified multilingual dataset of globally contested topics covering 18 languages, and propose ROOTS, a framework that profiles linguistic, perspective, and framing diversity. We evaluate a diverse set of eight models covering both open-source and proprietary families at varying scales, establishing that LLMs hedge their answers in English but commit to positions in lower-resource languages. Larger models maintain a more consistent cross-lingual worldview, while smaller models are more easily swayed. All models also exhibit high sensitivity to minor surface-level input variations, producing notable stance even in larger models.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.34278 [cs.CL] |
| (or arXiv:2610.34278v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.34278 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Wiem Ben Rim [view email]
[v1]
Sun, 25 Oct 2026 13:14:41 UTC (287 KB)
来源：arXiv cs.CL · arxiv.org