Making Corgis Important for Honeycomb Classification: Adversarial Attacks on Concept-based Explainability Tools

Open in new window