This repository provides the algorithm demonstration for C3A(-E) proposed in IEEE TNNLS 2026 submission 📃 Seeing What Few-shot Learners See: Contrastive Cross-Class Attribution for Explainability.