mlsanityData to 7 Oct 2026

Can Finetuing LLMs on Small Human Samples Increase Heterogeneity, Alignment, and Belief-Action Coherence?

arXiv 2511.21218 ↗First posted 26 Nov 2025cs.CLcited by 2

Not on the map. Why?

Abstract (arXiv)

There is ongoing debate about whether large language models (LLMs) can serve as substitutes for human participants in survey and experimental research. While recent work in fields such as marketing and psychology has explored the potential of LLM-based simulation, a growing body of evidence cautions against this practice: LLMs often fail to align with real human behavior, exhibiting limited diversity, systematic misalignment for minority subgroups, insufficient within-group variance, and discrepancies between stated beliefs and actions. This study examines an important and distinct question in this domain: whether fine-tuning on a small subset of human survey data, such as that obtainable from a pilot study, can mitigate these issues and yield realistic simulated outcomes. Using a behavioral experiment on information disclosure, we compare human and LLM-generated responses across multiple dimensions, including distributional divergence, subgroup alignment, belief-action coherence, and the recovery of regression coefficients. We find that fine-tuning on small human samples substantially improves heterogeneity, alignment, and belief-action coherence relative to the base model. However, even the best-performing fine-tuned models fail to reproduce the regression coefficients of the original study, suggesting that LLM-generated data remain unsuitable for replacing human participants in formal inferential analyses.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 23 of its 37 references that are in the map. First posted on arXiv 26 Nov 2025.

Builds on

23 of 37 references.

References · 23

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
  3. 2020Extracting Training Data from Large Language Modelscited by 2,144
  4. 2022Out of One, Many: Using Language Models to Simulate Human Samplescited by 737
  5. 2023Can Large Language Models Be an Alternative to Human Evaluations?cited by 729
  6. 2022Using Large Language Models to Simulate Multiple Humans and Replicate Human Subject Studiescited by 506
  7. 2023Better Zero-Shot Reasoning with Role-Play Promptingcited by 220
  8. 2023Do LLMs exhibit human-like response biases? A case study in survey designcited by 176
  9. 2023AnnoLLM: Making Large Language Models to Be Better Crowdsourced Annotatorscited by 171
  10. 2024Large language models that replace human participants can harmfully misportray and flatten identity groupscited by 137
  11. 2025LLM Social Simulations Are a Promising Research Methodcited by 97
  12. 2025Language Model Fine-Tuning on Scaled Survey Data for Predicting Distributions of Public Opinionscited by 64
11 more
  1. 2024Large Language Models Assume People are More Rational than We Really arecited by 49
  2. 2025Mind the Value-Action Gap: Do LLMs Act in Alignment with Their Values?cited by 49
  3. 2025The Prompt Makes the Person(a): A Systematic Evaluation of Sociodemographic Persona Prompting for Large Language Modelscited by 40
  4. 2023Sociodemographic Prompting is Not Yet an Effective Approach for Simulating Subjective Judgments with LLMscited by 36
  5. 2024Persona is a Double-edged Sword: Mitigating the Negative Impact of Role-playing Prompts in Zero-shot Reasoning Taskscited by 28
  6. 2024Six Fallacies in Substituting Large Language Models for Human Participantscited by 24
  7. 2024Modeling Human Subjectivity in LLMs Using Explicit and Implicit Human Factors in Personascited by 23
  8. 2024Investigating the Impact of Data Contamination of Large Language Models in Text-to-SQL Translationcited by 21
  9. 2025Large Language Models Do Not Simulate Human Psychologycited by 17
  10. 2025Valid Survey Simulations with Limited Human Data: The Roles of Prompting, Fine-Tuning, and Rectificationcited by 17
  11. 2025Improving LLM Reasoning through Interpretable Role-Playing Steeringcited by 11

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Value-Action Alignment in Large Language Models under Privacy-Prosocial Conflictcited by 0
  2. 2026When Can LLM Digital Twins Reduce Human Measurement? From Behavioral Fidelity to Statistical Substitutabilitycited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Evaluating LLMs as Human Surrogates in Controlled Experiments
  2. 2023The Challenge of Using LLMs to Simulate Human Behavior: A Causal Inference Perspective
  3. 2026Beyond the Mean: Three-Axis Fidelity for Aligning LLM-Based Survey Simulators from Small Pilot Data
  4. 2024One fish, two fish, but not the whole sea: Alignment reduces language models' conceptual diversity
  5. 2024Systematic Biases in LLM Simulations of Debates
  6. 2025SimBench: Benchmarking the Ability of Large Language Models to Simulate Human Behaviors
  7. 2025In Silico Development of Psychometric Scales: Feasibility of Representative Population Data Simulation with LLMs
  8. 2025Are LLM Agents Behaviorally Coherent? Latent Profiles for Social Simulation