Pretrained Persona Mixture Models and Tandem Models for Human Simulation
cs.CL
Submitted: 2026-09-18
Updated: 2026-10-07
Comments: 11 pages in body, 36 with appendices. 7 figures. 8 tables
License: http://creativecommons.org/licenses/by/4.0/
The gist: We argue here that the current dominant practice in LLM human simulation: prompting instruction-tuned assistant language models to role-play personas, is inaccurate and produces stereotyped
Terminology
Abstract
We argue here that the current dominant practice in LLM human simulation: prompting instruction-tuned assistant language models to role-play personas, is inaccurate and produces stereotyped predictions (lacking natural diversity). It has previously been shown that LLMs can be bound to personas using naturalistic, freetext dialog avoiding stereotyping. Here we show that binding can also be achieved using short, individual samples of dialog from specific people. Demographics can be added later without negative effects by simply querying the model. We use the term Persona Mixture Models (PMMs) for well-calibrated human models, currently realized as pretrained base models. We show that PMMs produce more accurate predictions than instruction-tuned models and retain more of the lexical, semantic, and pragmatic diversity found in human dialog. We measure realism and diversity of LLMs simulating human interlocutors across a diverse set of corpora spanning open-domain text, human-AI chat, and task-oriented dialogue between human speakers. However, base pretrained models can produce out-of-domain dialog and may lose some of the human's internal state over long contexts. We propose and explore tandem models which combine a pre-trained model with an instruction-tuned supervisor. Tandem models achieve the best overall accuracy and diversity in our experiments.
Sources
- Using Large Language Models to Simulate Multiple Humans and Replicate Human Subject Studies
- Can Large Language Models Transform Computational Social Science?
- Large Language Models as User-Agents for Evaluating Task-Oriented-Dialogue Systems
- Lost in Simulation: LLM-Simulated Users are Unreliable Proxies for Human Users in Agentic Evaluations
- $\tau$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains
- Mind the Sim2Real Gap in User Simulation for Agentic Tasks
- HumanLM: Simulating Users with State Alignment Beats Response Imitation
- Consistently Simulating Human Personas with Multi-Turn Reinforcement Learning
- Learning to Simulate Human Dialogue
- In-Context Learning User Simulators for Task-Oriented Dialog Systems
- LLM Agents Grounded in Self-Reports Enable General-Purpose Simulation of Individuals
- Deep Binding of Language Model Virtual Personas: a Study on Approximating Political Partisan Misperceptions
- Understanding the Effects of RLHF on LLM Generalisation and Diversity
- Does Writing with Language Models Reduce Content Diversity?
- The Price of Format: Diversity Collapse in LLMs
- Artificial Hivemind: The Open-Ended Homogeneity of Language Models (and Beyond)
- Flipping the Dialogue: Training and Evaluating User Language Models
- LLM Generated Persona is a Promise with a Catch
- Real or Robotic? Assessing Whether LLMs Accurately Simulate Qualities of Human Responses in Dialogue
- Moral Mimicry: Large Language Models Produce Moral Rationalizations Tailored to Political Identity
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering