"Evaluating NLP Models via Contrast Sets."

Matt Gardner et al. (2020)
a service of Schloss Dagstuhl - Leibniz Center for Informatics