Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2504.19445 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910920522858496 |
|---|---|
| author | Lu, Yi-Long Zhang, Chunhui Wang, Wei |
| author_facet | Lu, Yi-Long Zhang, Chunhui Wang, Wei |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_19445 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Systematic Bias in Large Language Models: Discrepant Response Patterns in Binary vs. Continuous Judgment Tasks Lu, Yi-Long Zhang, Chunhui Wang, Wei Computation and Language 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. |
| title | Systematic Bias in Large Language Models: Discrepant Response Patterns in Binary vs. Continuous Judgment Tasks |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2504.19445 |