arXiv cs.MA· Jinyi Ye, Lei Cao, Ding Chen, Emilio Ferrara·· 13 小时前AI 评分36
LLM 社会模拟中的蝴蝶效应：从架构敏感性到鲁棒性审计
The Butterfly Effect in LLM Social Simulations: From Architecture Sensitivity to Robustness Audits
AI 导读
研究通过重复囚徒困境和社交媒体回音室实验，在四个 LLM 系列上考察表示与设计层面的扰动如何改变集体结果。仅改变匹配人格的表示方式，GPT-5.2 的双人合作率就偏移 75.2 个百分点，且多智能体交互会放大或衰减这种敏感性。作者提出 TRAILS 分类框架与 Social Simulation Audit Card，用于系统审计这类依赖关系并校准科学结论的适用范围。
正文 · 原文
Abstract:A butterfly effect occurs when a local perturbation propagates through a system and produces divergence in its collective outcome. In LLM social simulations, this creates a particular challenge for scientific inference: sensitivity to implementation choices can enter agent interactions and alter the collective outcomes on which scientific claims are based. We study this problem through a repeated Prisoner's Dilemma and a social-media echo chamber, spanning representation- and design-level perturbations across four LLM families. We find that controlled perturbations can substantially alter collective outcomes, with sensitivity varying across both perturbations and models. In the Prisoner's Dilemma, changing only the representation of a matched persona shifts two-agent cooperation by 75.2 percentage points for GPT-5.2, while game framing and memory representation produce smaller or less consistent effects. Multi-agent interaction further reshapes representation sensitivity, producing amplification or attenuation depending on the model and perturbation. In the echo-chamber simulation, persona representation changes interaction targeting and is accompanied by divergence in collective interaction structure, extending representation sensitivity from agent behavior to network-level outcomes. These results show that robustness cannot be inferred from a single simulation configuration and that local sensitivity can change as it propagates through interaction. We introduce TRAILS (Taxonomy for Robustness Audits In LLM Simulations) and the Social Simulation Audit Card as a framework for systematically auditing these dependencies and calibrating the scope of scientific claims to the robustness evidence that supports them.
| Subjects: | Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2610.28981 [cs.MA] |
| (or arXiv:2610.28981v1 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2610.28981 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jinyi Ye [view email]
[v1]
Tue, 20 Oct 2026 23:13:27 UTC (3,302 KB)
来源：arXiv cs.MA · arxiv.org