Towards Transparent and Explainable Attention Models

ACL 2020