Bias

We study how models behave in the presence of the identified terms, and whether they contribute to biasing the performance of trained models.

Hate speech is in the eye of the beholder Exploring bias on hate perception