通过公平感知集成对文本安全分类器进行去偏
Debiasing Text Safety Classifiers through a Fairness-Aware Ensemble
AI总结：
本文提出一种公平感知集成后处理方法，通过结合阈值无关公平指标与公平数据重加权，在几乎不影响性能的前提下缓解闭源文本安全分类器的反事实偏见。
摘要：
大型语言模型（LLMs）日益广泛的使用要求高性能的防护机制来确保LLMs输入与输出的安全。当这些安全防护措施在不平衡数据上训练时，它们可能学习到社会偏见。我们提出一种轻量级的后处理方法，用于在闭源文本安全分类器中缓解反事实公平性问题。我们的方法涉及构建一个集成模型，该集成不仅优于输入分类器并使其与策略对齐，还充当去偏正则化器。我们引入两个与阈值无关的指标来评估模型的反事实公平性，并展示如何将这些指标与公平数据重加权（Fair Data Reweighting，FDW）相结合来帮助缓解偏见。我们创建了一个扩展的Open AI数据集，以及一个基于用户提示生成的新模板化LLM生成数据集，这两个数据集都在身份群体之间实现了反事实平衡，并覆盖四个关键安全领域；我们将致力于公开发布这些数据集。我们的结果表明，所提方法在提升反事实公平性的同时，对模型性能的影响极小。
英文摘要：
Increasing use of large language models (LLMs) demand performant guardrails to ensure the safety of inputs and outputs of LLMs. When these safeguards are trained on imbalanced data, they can learn the societal biases. We present a light-weight, post-processing method for mitigating counterfactual fairness in closed-source text safety classifiers. Our approach involves building an ensemble that not only outperforms the input classifiers and policy-aligns them, but also acts as a debiasing regularizer. We introduce two threshold-agnostic metrics to assess the counterfactual fairness of a model, and demonstrate how combining these metrics with Fair Data Reweighting (FDW) helps mitigate biases. We create an expanded Open AI dataset, and a new templated LLM-generated dataset based on user-prompts, both of which are counterfactually balanced across identity groups and cover four key areas of safety; we will work towards publicly releasing these datasets. Our results show that our approach improves counterfactual fairness with minimal impact on model performance.
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。
