Exploring attention weights in transformer-based models with linguistic knowledge. - View it on GitHub