用心理社会数据构建人物模型，让大模型更真实地模拟人类行为
The Need for a Socially-Grounded Persona Framework for User Simulation
- 基于141项心理社会问卷构建人物框架SCOPE
- 非人口统计类人物使行为预测准确率显著提升
- 适合需要高真实感社交模拟的研究与应用
合成人物常用于引导大语言模型进行社会模拟，但多数仍基于粗粒度的人口统计特征或摘要构建。本文提出一种基于社会心理学的框架SCOPE，其数据来自124名美国参与者完成的141项、历时两小时的心理社会协议。在7个模型中测试发现，仅使用人口统计信息的人物只能解释约1.5%的人类反应相似性方差；加入心理社会维度后，行为预测能力提升，过度强调现象减少；基于价值观与身份构建的非人口统计人物表现出强对齐性且偏差更低。该趋势在包含441个对齐问题的SimBench上得到验证，SCOPE人物优于默认提示和NVIDIA Nemotron人物，且可提升Nemotron人物表现。结果表明，人物质量取决于心理社会结构而非人口统计模板。
原文摘要 · Abstract (English)
Synthetic personas are widely used to condition large language models (LLMs) for social simulation, yet most personas are still constructed from coarse sociodemographic attributes or summaries. We revisit persona creation by introducing SCOPE, a socially grounded framework for persona construction and evaluation, built from a 141-item, two-hour sociopsychological protocol collected from 124 U.S.-based participants. Across seven models, we find that demographic-only personas are a structural bottleneck: demographics explain only ~1.5% of variance in human response similarity. Adding sociopsychological facets improves behavioral prediction and reduces over-accentuation, and non-demographic personas based on values and identity achieve strong alignment with substantially lower bias. These trends generalize to SimBench (441 aligned questions), where SCOPE personas outperform default prompting and NVIDIA Nemotron personas, and SCOPE augmentation improves Nemotron-based personas. Our results indicate that persona quality depends on sociopsychological structure rather than demographic templates or summaries.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。