Applications based on large language models (LLMs), such as multi-agent simulations, require population diversity among agents. We identify a pervasive failure mode we term \emph{Persona Collapse}: agents each assigned a distinct profile nonetheless converge into a narrow behavioral mode, producing a homogeneous simulated population. To quantify persona collapse, we propose a framework that measures how much of the persona space a population occupies (Coverage), how evenly agents spread across it (Uniformity), and how rich the resulting behavioral patterns are (Complexity). Evaluating ten LLMs on personality simulation (BFI-44), moral reasoning, and self-introduction, we observe persona collapse along two axes: (1) Dimensions: a model can appear diverse on one axis yet structurally degenerate on another, and (2) Domains: the same model may collapse the most in personality yet be the most diverse in moral reasoning. Furthermore, item-level diagnostics reveal that behavioral variation tracks coarse demographic stereotypes rather than the fine-grained individual differences specified in each persona. Counter-intuitively, \textbf{the models achieving the highest per-persona fidelity consistently produce the most stereotyped populations}. We release our toolkit and data to support population-level evaluation of LLMs.

翻译：基于大语言模型（LLM）的应用（如多智能体模拟）要求智能体具有种群多样性。我们发现一种普遍存在的失败模式，称之为“角色坍塌”：被分配不同设定档的智能体最终收敛到狭窄的行为模式，产生同质化的模拟种群。为量化角色坍塌，我们提出一个框架，通过测量种群占据角色空间的范围（覆盖率）、智能体在空间中分布的均匀程度（均匀性）以及由此产生的行为模式的丰富程度（复杂性）进行评估。在人格模拟（BFI-44）、道德推理和自我引介任务上对十个LLM进行评估后，我们观察到角色坍塌沿两个轴发生：（1）维度轴：模型在某一维度上可能表现多样，但在另一维度上结构退化；（2）领域轴：同一模型可能在人格模拟中最严重坍塌，却在道德推理中最多样化。此外，项目级诊断显示，行为变异追踪的是粗粒度的刻板印象，而非设定档中指定的细粒度个体差异。反直觉的是，**在单角色保真度上表现最佳的模型，始终生成最刻板化的种群**。我们发布相关工具包与数据，以支持对LLM进行种群级评估。