表面公平，深层偏见：语言模型中的偏见比较研究
Surface Fairness, Deep Bias: A Comparative Study of Bias in Language Models
AI总结：
该研究对比大语言模型的偏见度量方式，发现预提示人格的MMLU测试偏见差异微弱，打分任务和薪资建议任务偏见更显著，提示LLM个性化发展下的偏见风险。
摘要：
现代语言模型基于海量数据训练，这些数据不可避免地包含争议性与刻板印象内容，涉及性别、出身、年龄等各类偏见。因此模型会表达带有偏见的观点，或根据预设人格或用户人格产生不同结果。本文研究了大语言模型（LLM）偏见的多种代理度量指标。我们发现，在多学科基准MMLU（大规模多任务语言理解基准）上用预提示人格评估模型时，得分差异可忽略且大多随机。但如果重新设计任务，要求模型为用户的答案打分，则会显现更显著的偏见迹象。最后，当向模型咨询薪资谈判建议时，答案中存在明显偏见。随着近期LLM助手记忆与个性化的发展趋势，这些问题从另一角度凸显：现代LLM用户无需预提示自身人格描述，因为模型已掌握其社会人口统计特征。
英文摘要：
Modern language models are trained on large amounts of data. These data inevitably include controversial and stereotypical content, which contains all sorts of biases related to gender, origin, age, etc. As a result, the models express biased points of view or produce different results based on the assigned personality or the personality of the user. In this paper, we investigate various proxy measures of bias in large language models (LLMs). We find that evaluating models with pre-prompted personae on a multi-subject benchmark (MMLU) leads to negligible and mostly random differences in scores. However, if we reformulate the task and ask a model to grade the user's answer, this shows more significant signs of bias. Finally, if we ask the model for salary negotiation advice, we see pronounced bias in the answers. With the recent trend for LLM assistant memory and personalization, these problems open up from a different angle: modern LLM users do not need to pre-prompt the description of their persona since the model already knows their socio-demographics.
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。
