Interpretability Illusions in the Generalization of Simplified Models

Open in new window