Evaluating the Stability of Semantic Concept Representations in CNNs for Robust Explainability

Open in new window