Stable Personas: Dual-Assessment of Temporal Stability in LLM-Based Human Simulation

2026-01-30

AI Summary

This study addresses the challenge of inconsistent personality expression in large language models (LLMs) during human-like behavioral simulation, which undermines their credibility in social behavior research. The authors propose a dual-perspective framework integrating self-report and observer-based assessments, revealing for the first time that while LLMs exhibit highly stable self-reported personalities, their externally observable personality expressions significantly decay over extended dialogues. Drawing on 3,473 cross-session and 1,370 within-session interactions across seven mainstream LLMs, three semantically equivalent prompts, and four ADHD-related personality profiles, the work systematically identifies boundary conditions under which personality expression deteriorates. These findings provide critical empirical grounding for developing more reliable LLM-based social simulations.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Large Language Models (LLMs) acting as artificial agents offer the potential for scalable behavioral research, yet their validity depends on whether LLMs can maintain stable personas across extended conversations. We address this point using a dual-assessment framework measuring both self-reported characteristics and observer-rated persona expression. Across two experiments testing four persona conditions (default, high, moderate, and low ADHD presentations), seven LLMs, and three semantically equivalent persona prompts, we examine between-conversation stability (3,473 conversations) and within-conversation stability (1,370 conversations and 18 turns). Self-reports remain highly stable both between and within conversations. However, observer ratings reveal a tendency for persona expressions to decline during extended conversations. These findings suggest that persona-instructed LLMs produce stable, persona-aligned self-reports, an important prerequisite for behavioral research, while identifying this regression tendency as a boundary condition for multi-agent social simulation.

Problem

• persona stability
• large language models
• behavioral simulation
• temporal consistency
• human-like agents

Innovation

• temporal stability
• persona consistency
• dual-assessment framework
• LLM-based human simulation
• behavioral research