LLM智能体社会中的局部可预测性与集体保真度
Local Predictability and Collective Fidelity in LLM-Agent Societies
浏览论文内容
中文总结 AI 辅助
本研究通过比较个体与集体预测，验证了紧凑替代模型在LLM智能体社会中的有效性，发现邻居信息提升预测性能，并强调直接集体验证的必要性。
中文摘要 AI 辅助
紧凑的替代模型可以降低模拟大型语言模型社会的成本，但必须再现集体行为。我们使用9,455条已发表的轨迹和关于观点动态的新实验，比较了个体预测和集体预测。在所有16个公共数据设置中，邻居信息改善了个体预测，并在保留问题上改善了汇总的集体预测，尽管集体收益取决于迁移条件。对24条新陈述的测试并未证实早期从初始状态预测中观察到的对比历史效应。Qwen在观察三轮后受益于历史信息。这些发现促使进行直接的集体验证、对可用观测的明确限制，以及与简单基线的比较。
英文摘要
Compact surrogates could reduce the cost of simulating large language model societies, but must reproduce collective behavior. We compare individual predictions and collective forecasts using 9,455 published trajectories and new experiments on opinion dynamics. Neighbor information improves individual prediction in all 16 public-data settings and pooled collective forecasts on held-out questions, although collective gains depend on transfer conditions. Tests on 24 new statements do not confirm earlier contrasting history effects in forecasts from the initial state. Qwen benefits from history after three observed rounds. These findings motivate direct collective validation, explicit limits on available observations, and comparisons with simple baselines.
