Large Language Models Often Say One Thing and Do Another
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909532527001600 |
|---|---|
| author | Xu, Ruoxi Lin, Hongyu Han, Xianpei Zheng, Jia Zhou, Weixiang Sun, Le Sun, Yingfei |
| author_facet | Xu, Ruoxi Lin, Hongyu Han, Xianpei Zheng, Jia Zhou, Weixiang Sun, Le Sun, Yingfei |
| contents | As large language models (LLMs) increasingly become central to various applications and interact with diverse user populations, ensuring their reliable and consistent performance is becoming more important. This paper explores a critical issue in assessing the reliability of LLMs: the consistency between their words and deeds. To quantitatively explore this consistency, we developed a novel evaluation benchmark called the Words and Deeds Consistency Test (WDCT). The benchmark establishes a strict correspondence between word-based and deed-based questions across different domains, including opinion vs. action, non-ethical value vs. action, ethical value vs. action, and theory vs. application. The evaluation results reveal a widespread inconsistency between words and deeds across different LLMs and domains. Subsequently, we conducted experiments with either word alignment or deed alignment to observe their impact on the other aspect. The experimental results indicate that alignment only on words or deeds poorly and unpredictably influences the other aspect. This supports our hypothesis that the underlying knowledge guiding LLMs' word or deed choices is not contained within a unified space. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_07003 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Large Language Models Often Say One Thing and Do Another Xu, Ruoxi Lin, Hongyu Han, Xianpei Zheng, Jia Zhou, Weixiang Sun, Le Sun, Yingfei Computation and Language As large language models (LLMs) increasingly become central to various applications and interact with diverse user populations, ensuring their reliable and consistent performance is becoming more important. This paper explores a critical issue in assessing the reliability of LLMs: the consistency between their words and deeds. To quantitatively explore this consistency, we developed a novel evaluation benchmark called the Words and Deeds Consistency Test (WDCT). The benchmark establishes a strict correspondence between word-based and deed-based questions across different domains, including opinion vs. action, non-ethical value vs. action, ethical value vs. action, and theory vs. application. The evaluation results reveal a widespread inconsistency between words and deeds across different LLMs and domains. Subsequently, we conducted experiments with either word alignment or deed alignment to observe their impact on the other aspect. The experimental results indicate that alignment only on words or deeds poorly and unpredictably influences the other aspect. This supports our hypothesis that the underlying knowledge guiding LLMs' word or deed choices is not contained within a unified space. |
| title | Large Language Models Often Say One Thing and Do Another |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2503.07003 |