Persona Generators: Generating Diverse Synthetic Personas at Scale

TL;DR

This paper introduces Persona Generators, a novel method using large language models and iterative optimization to produce diverse synthetic personas that better cover the range of possible human behaviors, especially rare traits.

Contribution

It presents a new approach for generating diverse synthetic populations with high coverage using iterative refinement and LLM-based mutation, improving over existing methods.

Findings

Generated populations outperform baselines on six diversity metrics.

Evolved generators produce rare trait combinations more effectively.

Method scales to arbitrary contexts and enhances diversity coverage.

Abstract

Evaluating AI systems that interact with humans requires understanding their behavior across diverse user populations, but collecting representative human data is often expensive or infeasible, particularly for novel technologies or hypothetical future scenarios. Recent work in Generative Agent-Based Modeling has shown that large language models can simulate human-like synthetic personas with high fidelity, accurately reproducing the beliefs and behaviors of specific individuals. However, most approaches require detailed data about target populations and often prioritize density matching (replicating what is most probable) rather than support coverage (spanning what is possible), leaving long-tail behaviors underexplored. We introduce Persona Generators, functions that can produce diverse synthetic populations tailored to arbitrary contexts. We apply an iterative improvement loop based on AlphaEvolve, using large language models as mutation operators to refine our Persona Generator code over hundreds of iterations. The optimization process produces lightweight Persona Generators that can automatically expand small descriptions into populations of diverse synthetic personas that maximize coverage of opinions and preferences along relevant diversity axes. We demonstrate that evolved generators substantially outperform existing baselines across six diversity metrics on held-out contexts, producing populations that span rare trait combinations difficult to achieve in standard LLM outputs.