Bias Against 93 Stigmatized Groups in Masked Language Models and Downstream Sentiment Classification Tasks

Open in new window