Acquiescence Bias in Large Language Models

Daniel Braun
9/10/2025
cs.CL

No official code found. Best implementation: Responsible-NLP/Acquiescence-Bias-in-Large-Language-Models (0 stars, last commit 2025-09-12).

Abstract

Acquiescence bias, i.e. the tendency of humans to agree with statements in surveys, independent of their actual beliefs, is well researched and documented. Since Large Language Models (LLMs) have been shown to be very influenceable by relatively small changes in input and are trained on human-generated data, it is reasonable to assume that they could show a similar tendency. We present a study investigating the presence of acquiescence bias in LLMs across different models, tasks, and languages (English, German, and Polish). Our results indicate that, contrary to humans, LLMs display a bias towards answering no, regardless of whether it indicates agreement or disagreement.

View on arXivView PDF

Code Implementations(1)

Responsible-NLP/Acquiescence-Bias-in-Large-Language-ModelsCommunity72%

Acquiescence Bias in Large Language Models

00PythonSep 5, 20251 years ago

Similar Papers

Prompt Perturbations Reveal Human-Like Biases in Large Language Model Survey Responses

85% match2025

Conformity in Large Language Models

85% match2024

Systematic Bias in Large Language Models: Discrepant Response Patterns in Binary vs. Continuous Judgment Tasks

85% match2025

A Systematic Analysis of Biases in Large Language Models

84% match2025

Surface Fairness, Deep Bias: A Comparative Study of Bias in Language Models

84% match2025

Investigating the Effects of Cognitive Biases in Prompts on Large Language Model Outputs

83% match2025

Cite this paper

@article{braun2025acquiescence,
  title  = {Acquiescence Bias in Large Language Models},
  author = {Daniel Braun},
  year   = {2025},
  eprint = {2509.08480},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2509.08480v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2509.08480.svg)](https://codeofpaper.com/paper/2509.08480)

Discussion