arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2407.07329cs.CL
收藏

分化概率揭示了GPT-4中同质性偏差的脆弱性

Probability of Differentiation Reveals Brittleness of Homogeneity Bias in GPT-4

Washington University in St. Louis(圣路易斯华盛顿大学);Rutgers University(罗格斯大学)

Messi H. J. Lee, Calvin K. Lai

AI总结:

本研究通过提示GPT-4生成18个情境线索相关的补全并以分化概率分析变异性,发现过往观测的大语言模型同质性偏差或源于编码器模型,且该偏差本身具有脆弱性,微小提示改动即可显著改变其表现。

摘要:

大语言模型(LLM)中的同质性偏差指的是其倾向于将某些群体的表征同质化,程度超过其他群体。此前记录该偏差的研究主要使用编码器模型,这可能无意中引入了偏差。为解决这一局限,我们提示GPT-4生成与18个情境线索(即影响个体感知情境的环境中特定、可测量的元素)相关的单个单词/短语补全,并使用分化概率比较这些补全的变异性。该方法直接从模型输出评估同质性偏差,绕过了编码器模型。在五项研究中,我们发现同质性偏差在不同情境线索和写作提示下波动极大,表明过往研究中观察到的偏差可能反映的是编码器模型内部的偏差,而非LLM的偏差。此外,我们发现LLM中的同质性偏差很脆弱,因为即使是提示中微小、任意的改动也能显著改变偏差的表现形式。未来研究应进一步探索长文本生成中句法特征和主题选择的变化如何影响LLM的同质性偏差。

英文摘要:

Homogeneity bias in Large Language Models (LLMs) refers to their tendency to homogenize the representations of some groups compared to others. Previous studies documenting this bias have predominantly used encoder models, which may have inadvertently introduced biases. To address this limitation, we prompted GPT-4 to generate single word/expression completions associated with 18 situation cues-specific, measurable elements of environments that influence how individuals perceive situations and compared the variability of these completions using probability of differentiation. This approach directly assessed homogeneity bias from the model's outputs, bypassing encoder models. Across five studies, we find that homogeneity bias is highly volatile across situation cues and writing prompts, suggesting that the bias observed in past work may reflect those within encoder models rather than LLMs. Furthermore, we find that homogeneity bias in LLMs is brittle, as even minor and arbitrary changes in prompts can significantly alter the expression of biases. Future work should further explore how variations in syntactic features and topic choices in longer text generations influence homogeneity bias in LLMs.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑