XAM: Interactive Explainability for Authorship Attribution Models