评估大型语言模型（LLMs）中从抽象刻板印象到具体社会决策的区域偏差
Evaluating Regional Bias in LLMs From Abstract Stereotype to Concrete Social Decision-Making
浏览论文内容
中文总结 AI 辅助
该研究提出S2D框架，评估6个LLMs在34个中国省级行政区的区域偏差，发现其普遍存在且具系统性，推动相关评估与缓解。
中文摘要 AI 辅助
大型语言模型（LLMs）的区域偏差可能会影响对区域群体的认知以及对来自不同区域个体的决策。然而现有研究通常分别考察这些表现，导致其结构与后果尚不明确。我们提出Stereotypes-to-Decisions（S2D）这一系统框架，用于评估从抽象刻板印象到具体社会决策的区域偏差。S2D覆盖中国全部34个省级行政区，通过温暖度（感知的友好度与可信度）和能力度（感知的能力与智力）的刻板印象评分，以及教育、职业、社会互动领域的配对选择任务，对6个LLMs进行评估。结果显示区域评分存在显著的区域差异，不同模型间存在相当的一致性，尤其在能力度和职业决策方面。此外，这些模式与区域经济和数字发展指标相关，呈现出混合的类人刻板印象，部分区域在某一维度评分高而另一维度低。它们在中文和英文提示下基本保持稳定。总体而言，我们的发现表明LLMs中的区域偏差普遍存在、具有系统性且会产生后果，推动更具区域意识的评估与缓解工作。
英文摘要
Regional bias in large language models (LLMs) may shape both perceptions of regional groups and decisions about individuals from different regions. Yet existing studies often examine these manifestations separately, leaving their structure and consequences unclear. We introduce Stereotypes-to-Decisions (S2D), a systematic framework evaluating regional bias from abstract stereotypes to concrete social decisions. Covering all 34 provincial-level administrative regions of China, S2D evaluates six LLMs using stereotype ratings of Warmth (perceived friendliness and trustworthiness) and Competence (perceived capability and intelligence), along with paired-choice tasks across Education, Occupation, and Social Interaction. Results reveal substantial regional differences in regional scores, with considerable agreement across models, especially for Competence and Occupation decisions. Furthermore, these patterns are associated with regional economic and digital development indicators and display mixed human-like stereotypes, with some regions rated highly on one dimension but poorly on the other. They also remain largely stable across Chinese and English prompts. Overall, our findings show that regional bias in LLMs is prevalent, systematic, and consequential, motivating more regionally aware evaluation and mitigation.
