PsyAgent: Constructing Human-like Agents Based on Psychological Modeling and Contextual Interaction

TL;DR

PsyAgent is a framework for creating human-like agents that combine psychological traits with contextual social norms, enabling stable yet adaptable behavior through a structured, data-efficient approach.

Contribution

It introduces a schema-first, trait and context-based method for synthesizing and fine-tuning human-like agents, improving trait-faithfulness and stability.

Findings

Enhanced trait-faithfulness in agent behavior

Improved long-horizon stability

Competitive performance with larger instruction-tuned models

Abstract

Human-like agents must express stable dispositions while adapting to roles, relationships, and norms. We present PsyAgent, a schema-first framework that operationalizes the trait-context interface by coupling a Big Five trait prior with explicit social-structural conditioning. PsyAgent comprises (i) Individual Structure (IS), a machine-usable trait-grounded profile, and (ii) Multi-Scenario Contexting (MSC), a curated library of role-relationship-norm frames spanning eight everyday arenas. At inference, fixed structured prompts couple the active MSC frame with the IS profile, encouraging behavior that is stable yet context-sensitive. To demonstrate learnability beyond prompt engineering, we use IS and MSC to synthesize supervision and fine-tune compact backbones with PEFT (SFT and optional DPO). Under a controlled psychometric-style evaluation protocol in percentile space, PsyAgent improves trait-faithfulness and long-horizon stability, and is competitive with several larger general-purpose instruction-tuned baselines under matched decoding and scoring controls. We further triangulate the automatic protocol with external benchmarks and a small blinded human study. Overall, PsyAgent provides a precise and data-efficient approach to personality-grounded, norm-aware agents.