DeepPersona: A Generative Engine for Scaling Deep Synthetic Personas

TL;DR

DeepPersona presents a taxonomy-guided, two-stage engine to generate narrative-complete synthetic personas at scale. By constructing an extensive Human-Attribute Tree from real dialogues and employing progressive sampling anchored by core traits, it achieves unprecedented depth and diversity in profiles. Intrinsic metrics show richer attribute coverage and uniqueness, while extrinsic tests demonstrate improved LLM personalization, more faithful social simulations, and better alignment with Big Five distributions. The framework offers a privacy-preserving, controllable toolkit for research in agent-based modeling, personalization, and human-AI alignment.

Abstract

Simulating human profiles by instilling personas into large language models (LLMs) is rapidly transforming research in agentic behavioral simulation, LLM personalization, and human-AI alignment. However, most existing synthetic personas remain shallow and simplistic, capturing minimal attributes and failing to reflect the rich complexity and diversity of real human identities. We introduce DEEPPERSONA, a scalable generative engine for synthesizing narrative-complete synthetic personas through a two-stage, taxonomy-guided method. First, we algorithmically construct the largest-ever human-attribute taxonomy, comprising over hundreds of hierarchically organized attributes, by mining thousands of real user-ChatGPT conversations. Second, we progressively sample attributes from this taxonomy, conditionally generating coherent and realistic personas that average hundreds of structured attributes and roughly 1 MB of narrative text, two orders of magnitude deeper than prior works. Intrinsic evaluations confirm significant improvements in attribute diversity (32 percent higher coverage) and profile uniqueness (44 percent greater) compared to state-of-the-art baselines. Extrinsically, our personas enhance GPT-4.1-mini's personalized question answering accuracy by 11.6 percent on average across ten metrics and substantially narrow (by 31.7 percent) the gap between simulated LLM citizens and authentic human responses in social surveys. Our generated national citizens reduced the performance gap on the Big Five personality test by 17 percent relative to LLM-simulated citizens. DEEPPERSONA thus provides a rigorous, scalable, and privacy-free platform for high-fidelity human simulation and personalized AI research.

Figures (8)

• Figure 1: Current persona generation methods face a trade-off between quantity and depth. While approaches like PersonaHub ge2024scaling achieve massive scale with shallow depth, DeepPersona uniquely scales both, automatically enriching PersonaHub's billion profiles with hundreds of structured attributes.
• Figure 2: DeepPersona Overview. Stage 1 builds a comprehensive Human-Attribute Tree by mining self-disclosure QA (left) and merging semantically validated paths (middle). Stage 2 anchors core traits, samples tree nodes, and fills values via LLM, yielding a narrative-complete profile (right).
• Figure 3: This sunburst chart shows domain coverage for taxonomy generation. Segment sizes are proportional to domain share, highlighting a balanced distribution without a single dominant topic.
• Figure 4: Personalization Prompting Example
• Figure 5: Personalization Evaluation