Dynamic Interpretability for Model Comparison via Decision Rules