Mitigating social biases typically requires identifying the social groups
associated with each data sample. In this paper, we present DAFair, a novel
approach to address social bias in language models. Unlike traditional methods
that rely on explicit demographic labels, our approach does not require any
such information. Instead, we leverage predefined prototypical demographic
texts and incorporate a regularization term during the fine-tuning process to
mitigate bias in the model's representations. Our empirical results across two
tasks and two models demonstrate the effectiveness of our method compared to
previous approaches that do not rely on labeled data. Moreover, with limited
demographic-annotated data, our approach outperforms common debiasing
approaches.

DAFair 是一种用于减轻语言模型中社会偏见的新方法，通过预定义的典型人口统计文本，并在微调过程中加入正则化项来纠正模型的表示中的偏见，从而在两个任务和两个模型上实证结果显示了我们方法的有效性。此外，即使在有限的人口统计注释数据下，我们的方法也优于常见的去偏方法。