# Mitigating Social Desirability Bias in Random Silicon Sampling

arXiv: https://arxiv.org/abs/2512.22725 · HTML: https://mlsanity.com/paper/2512.22725/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 31 references that are in the map.
- First posted on arXiv 27 Dec 2025.

## Builds on (most cited first, 19 of 19)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [Generative Agents: Interactive Simulacra of Human Behavior](https://mlsanity.com/paper/2304.03442/) (2023)
- [Out of One, Many: Using Language Models to Simulate Human Samples](https://mlsanity.com/paper/2209.06899/) (2022)
- [Bias and Fairness in Large Language Models: A Survey](https://mlsanity.com/paper/2309.00770/) (2023)
- [A Prompt Pattern Catalog to Enhance Prompt Engineering with ChatGPT](https://mlsanity.com/paper/2302.11382/) (2023)
- [Towards Understanding and Mitigating Social Biases in Language Models](https://mlsanity.com/paper/2106.13219/) (2021)
- [S3: Social-network Simulation System with Large Language Model-Empowered Agents](https://mlsanity.com/paper/2307.14984/) (2023)
- [Large Pre-trained Language Models Contain Human-like Biases of What is Right and Wrong to Do](https://mlsanity.com/paper/2103.11790/) (2021)
- [Personality Traits in Large Language Models](https://mlsanity.com/paper/2307.00184/) (2023)
- [Agentic Misalignment: How LLMs Could Be Insider Threats](https://mlsanity.com/paper/2510.05179/) (2025)
- [OASIS: Open Agent Social Interaction Simulations with One Million Agents](https://mlsanity.com/paper/2411.11581/) (2024)
- [ChatGPT-4 Outperforms Experts and Crowd Workers in Annotating Political Twitter Messages with Zero-Shot Learning](https://mlsanity.com/paper/2304.06588/) (2023)
- [Random Silicon Sampling: Simulating Human Sub-Population Opinion Using a Large Language Model Based on Group-Level Demographic Information](https://mlsanity.com/paper/2402.18144/) (2024)
- [Interactive Agents: Simulating Counselor-Client Psychological Counseling via Role-Playing LLM-to-LLM Interactions](https://mlsanity.com/paper/2408.15787/) (2024)
- [ElectionSim: Massive Population Election Simulation Powered by Large Language Model Driven Agents](https://mlsanity.com/paper/2410.20746/) (2024)
- [Actions Speak Louder than Words: Agent Decisions Reveal Implicit Biases in Language Models](https://mlsanity.com/paper/2501.17420/) (2025)
- [Explicit vs. Implicit: Investigating Social Bias in Large Language Models through Self-Reflection](https://mlsanity.com/paper/2501.02295/) (2025)
- [Towards More Accurate US Presidential Election via Multi-step Reasoning with Large Language Models](https://mlsanity.com/paper/2411.03321/) (2024)

## Built on by (most cited first, 1 of 1)

- [Beyond the Mean: Three-Axis Fidelity for Aligning LLM-Based Survey Simulators from Small Pilot Data](https://mlsanity.com/paper/2606.28963/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) are increasingly used to simulate population responses, a method known as ``Silicon Sampling''. However, responses to socially sensitive questions frequently exhibit Social Desirability Bias (SDB), diverging from real human data toward socially acceptable answers. Existing studies on social desirability bias in LLM-based sampling remain limited. In this work, we investigate whether minimal, psychologically grounded prompt wording can mitigate this bias and improve alignment between silicon and human samples. We conducted a study using data from the American National Election Study (ANES) on three LLMs from two model families: the open-source Llama-3.1 series and GPT-4.1-mini. We first replicate a baseline silicon sampling study, confirming the persistent Social Desirability Bias. We then test four prompt-based mitigation methods: reformulated (neutral, third-person phrasing), reverse-coded (semantic inversion), and two meta-instructions, priming and preamble, respectively encouraging analytics and sincerity. Alignment with ANES is evaluated using Jensen-Shannon Divergence with bootstrap confidence intervals. Our results demonstrate that reformulated prompts most effectively improve alignment by reducing distribution concentration on socially acceptable answers and achieving distributions closer to ANES. Reverse-coding produced mixed results across eligible items, while the Priming and Preamble encouraged response uniformity and showed no systematic benefit for bias mitigation. Our findings validate the efficacy of prompt-based framing controls in mitigating inherent Social Desirability Bias in LLMs, providing a practical path toward more representative silicon samples.