CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Models
This provides a benchmark for evaluating social biases in language models, addressing harm from biased representations, though it is incremental as it focuses on measurement rather than mitigation.
The authors tackled the problem of measuring social biases in masked language models by introducing the CrowS-Pairs dataset, which contains 1508 examples covering nine bias types, and found that three widely-used models substantially favored stereotyping sentences across all categories.
Pretrained language models, especially masked language models (MLMs) have seen success across many NLP tasks. However, there is ample evidence that they use the cultural biases that are undoubtedly present in the corpora they are trained on, implicitly creating harm with biased representations. To measure some forms of social bias in language models against protected demographic groups in the US, we introduce the Crowdsourced Stereotype Pairs benchmark (CrowS-Pairs). CrowS-Pairs has 1508 examples that cover stereotypes dealing with nine types of bias, like race, religion, and age. In CrowS-Pairs a model is presented with two sentences: one that is more stereotyping and another that is less stereotyping. The data focuses on stereotypes about historically disadvantaged groups and contrasts them with advantaged groups. We find that all three of the widely-used MLMs we evaluate substantially favor sentences that express stereotypes in every category in CrowS-Pairs. As work on building less biased models advances, this dataset can be used as a benchmark to evaluate progress.