Interpretability is Harder in the Multiclass Setting: Axiomatic Interpretability for Multiclass Additive Models

Open in new window