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

2025/04/28 by LU Yi-long, Lu, Yi-Long, Chunhui Zhang +3 · 1 citation
Computer Science · Medicine · Social Sciences · #Artificial Intelligence in Healthcare and Education #Computation and Language (cs.CL) #Computational and Text Analysis Methods #Explainable Artificial Intelligence (XAI) #FOS: Computer and information sciences

paper · pdf · doi:10.48550/arxiv.2504.19445

openalex publication_date 2025/04/28 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Large Language Models (LLMs) are increasingly used in tasks such as psychological text analysis and decision-making in automated workflows. However, their reliability remains a concern due to potential biases inherited from their training process. In this study, we examine how different response format: binary versus continuous, may systematically influence LLMs' judgments. In a value statement judgments task and a text sentiment analysis task, we prompted LLMs to simulate human responses and tested both formats across several models, including both open-source and commercial models. Our findings revealed a consistent negative bias: LLMs were more likely to deliver "negative" judgments in binary formats compared to continuous ones. Control experiments further revealed that this pattern holds across both tasks. Our results highlight the importance of considering response format when applying LLMs to decision tasks, as small changes in task design can introduce systematic biases.

Citations

• LLM Agents Grounded in Self-Reports Enable General-Purpose Simulation of Individuals
• Are Language Model Logits Calibrated?
• Are Large Language Models Consistent over Value-laden Questions?
• Testing theory of mind in large language models and humans
• Determinants of LLM-assisted Decision-Making
• Cognitive Bias in Decision-Making with LLMs
• Generative Language Models Exhibit Social Identity Biases
• Large Language Models Are Not Robust Multiple Choice Selectors
• Cognitive Architectures for Language Agents
• Large Language Models are not Fair Evaluators
• GPT-4 Technical Report

Cited by

• Measuring Cognitive Engagement in Collaborative Discourse with an Extended ICAP Framework: Comparing Human Annotation, In-Context Learning, and Reflective LLM Agents

Related

• Cognitive Biases in Large Language Models: A Survey and Mitigation Experiments
• The Bias is in the Details: An Assessment of Cognitive Bias in LLMs
• Benchmarking Cognitive Biases in Large Language Models as Evaluators
• Verbosity Bias in Preference Labeling by Large Language Models
• Investigating the Effects of Cognitive Biases in Prompts on Large Language Model Outputs