失败发生在漂移之前：LLM智能体社会中的价值观社会动态

中文摘要 AI 辅助

基于大型语言模型（LLM）的智能体越来越多地被用作社会科学研究中人类参与者的代理，然而它们能否忠实地模拟多样且冲突的人类价值体系仍不清楚。我们提出了一个以世界价值观调查（WVS）为基础的模拟框架，其中具有不同沟通风格的文化多样智能体参与纵向的、充满价值取向的讨论。在涉及1200个人设、15个主题和三个模型（GPT-4o、Gemini-2.5-Flash和Gemma-4-E4B）的约4000次对话中，我们评估了价值忠实度、价值漂移和对话真实性。我们发现，超过50%的人设从一开始就未能表达其被分配的WVS档案，而2-7%的人设在重复对话后发生漂移。去除人口统计细节的消融实验提高了某些模型的忠实度，但并未改变总体趋势：模拟的价值分布仍然系统性地偏离被分配的WVS档案。与人类讨论相比，模拟对话在风格一致性和语义多样性之间表现出不同的权衡，往往产生内容上多样但风格上重复的交流。这些发现表明，当前的LLM智能体能够生成看似合理的对话，但在代表和长期保持多样化人类价值档案方面仍是有限的代理。

英文摘要

Large Language Model (LLM)-based agents are increasingly used as proxies for human participants in social science research, yet it remains unclear whether they can faithfully simulate diverse and conflicting human value systems. We present a World Values Survey (WVS)-grounded simulation framework where culturally diverse agents with different communication styles engage in longitudinal, value-laden discussions. Across approximately 4,000 conversations involving 1,200 personas, 15 topics, and three models (GPT-4o, Gemini-2.5-Flash, and Gemma-4-E4B), we evaluate value faithfulness, value drift, and conversational realism. We find that more than 50\% of personas fail to express their assigned WVS profiles from the outset, while 2-7\% drift after repeated conversations. Ablations removing demographic details improve faithfulness for some models but do not change the broader trend: simulated value distributions still systematically deviate from the assigned WVS profiles. Compared to human discussions, simulated dialogues show a different trade-off between stylistic consistency and semantic diversity, often producing content-wise varied but stylistically repetitive exchanges. These findings suggest that current LLM agents can generate plausible conversations, but remain limited proxies for representing and preserving diverse human value profiles over time.

发表机构

• Mohamed Bin Zayed University of Artificial Intelligence（穆罕默德·本·扎耶德人工智能大学）

机构由 AI 辅助整理，请以论文原文为准。