发表机构
The International Joint Institute of Tianjin University; Tianjin University; TJUNLP Lab, School of Computer Science and Technology, Tianjin University; National Governance Institute, Tianjin Normal University; College of Computer and Information Engineering, Tianjin Normal University（天津大学国际联合研究院; 天津大学; 天津大学计算机科学与技术学院; 天津师范大学国家治理研究院; 天津师范大学计算机与信息工程学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究针对大型语言模型社会模拟中个体价值观建模的不足，提出ExpertIVS框架，通过14个社会学专家智能体重构个体画像，经多国家多人群实验验证其在价值观还原、泛化等方面性能优于基线。
AI 中文摘要
大型语言模型（LLM）智能体在社会模拟中展现出巨大潜力，但难以准确建模个体价值观体系。现有多数方法机械地将调查回复拼接为提示，存在语义碎片化问题，无法捕捉人类价值观体系的内在一致性；且通常采用静态选择题评估LLM的价值观体系，无法评估现实对话交互中的价值取向。为解决这些问题，我们提出ExpertIVS框架，该框架利用14个社会学专家智能体，通过结构化专业视角解读世界价值观调查（WVS）的回复，而非直接拼接回复。这些专家智能体执行深度语义重构，生成稳健且内部一致的个体画像。为评估动态交互中LLM与个体价值观体系的一致性，我们进一步引入多智能体辩论机制。对来自12个国家的480名个体开展的大量实验表明，ExpertIVS实现了90.78%的价值观还原保真度，且在价值观泛化方面显著优于基线方法（提升5.3%）。此外，ExpertIVS表现出强大的人格区分度和行为一致性，实现了从单纯回复拼接到真正社会学角色扮演的转变。
英文摘要
Large Language Model (LLM) agents have demonstrated considerable potential for social simulation, yet struggle to accurately model individual value systems. Most existing methods mechanically stitch survey responses into prompts, which suffer from semantic fragmentation, failing to capture the internal coherence of human value systems. The value systems of LLMs are typically assessed using static multiple-choice questions, which fail to evaluate the value orientation in real-world dialogue interactions. To address these issues, we propose ExpertIVS, a framework employing 14 Sociological Expert Agents to interpret World Values Survey (WVS) responses through structured professional perspectives, rather than direct responses concatenation. These expert agents perform deep semantic reconstruction to generate robust and internally consistent individual profiles. To evaluate the consistency between LLMs and individual value systems during dynamic interactions, we further introduce a multi-agent debate mechanism. Extensive experiments across 480 individuals from 12 countries demonstrate that ExpertIVS achieves 90.78% value restoration fidelity and significantly outperforms baselines in value generalization (+5.3%). Moreover, ExpertIVS exhibits strong personality discriminability and behavioral consistency, enabling a shift from mere response concatenation to genuine sociological role-playing.
