Assessing Social and Intersectional Biases in Contextualized Word Representations

Open in new window