Boosting Explainability through Selective Rationalization in Pre-trained Language Models

Open in new window