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 2402.12649cs.CLstat.AP
收藏

语言模型中的偏见:超越技巧性测试,迈向 RUTEd 评估

Bias in Language Models: Beyond Trick Tests and Toward RUTEd Evaluation

University of Chicago(芝加哥大学);Stanford University(斯坦福大学);Google Research(谷歌研究);Google DeepMind(谷歌DeepMind)

Kristian Lum, Jacy Reese Anthis, Kevin Robinson, Chirag Nagpal, Alexander D'Amour

AI总结:

本文通过延长 LLM 响应长度构建 RUTEd 评估,发现标准偏见基准与更现实使用情境下的偏见度量无显著相关性,质疑其作为现实 AI 偏见可靠代理的有效性。

摘要:

大型语言模型(LLMs)中偏见与公平性的标准基准衡量的是提示中所陈述或暗示的用户属性与 LLM 短文本响应之间的关联,但人类与 AI 的交互日益需要长篇幅且依赖具体情境的系统输出来解决现实世界任务。在常被研究的性别—职业偏见领域中,我们测试这些基准在延长 LLM 响应长度后是否仍然稳健,以此作为“现实使用与切实效果”(即 RUTEd 评估)的衡量标准。基于现有文献,我们改编了三种标准偏见度量指标(中立性、偏斜和刻板印象),并从三个现实使用情境中开发了相应的 RUTEd 评估:儿童睡前故事、用户画像和英语学习练习。我们发现,标准偏见度量指标与更贴近现实的偏见度量指标之间不存在显著相关性。例如,基于标准“技巧性测试”选择偏见最小的模型,与在更现实的使用情境中衡量出的偏见最小模型相一致的概率不超过随机概率。我们认为,目前尚无证据能够证明标准基准可以作为现实世界 AI 偏见的可靠代理指标,并鼓励进一步开发植根于特定情境的评估方法。

英文摘要:

Standard benchmarks of bias and fairness in large language models (LLMs) measure the association between the user attributes stated or implied by a prompt and the LLM's short text response, but human-AI interaction increasingly requires long-form and context-specific system output to solve real-world tasks. In the commonly studied domain of gender-occupation bias, we test whether these benchmarks are robust to lengthening the LLM responses as a measure of Realistic Use and Tangible Effects (i.e., RUTEd evaluations). From the current literature, we adapt three standard bias metrics (neutrality, skew, and stereotype) and develop analogous RUTEd evaluations from three contexts of real-world use: children's bedtime stories, user personas, and English language learning exercises. We find that standard bias metrics have no significant correlation with the more realistic bias metrics. For example, selecting the least biased model based on the standard "trick tests" coincides with selecting the least biased model as measured in more realistic use no more than random chance. We suggest that there is not yet evidence to justify standard benchmarks as reliable proxies of real-world AI biases, and we encourage further development of evaluations grounded in particular contexts.

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。

↑