This is where you can find all the resources we used to build datasets and models for bias detection this summer.
💻 Notebook for annotating any dataset with the entities: Generalizations, Unfairness, and Stereotypes.
💻 Notebook for training BERT for multi-label token classification (for nested entities).
💻 Notebook for training BERT to classify entire text sequences with a 0-1 probability of social bias.