Code for the ICLR 2022 paper "Attention-based interpretability with Concept Transformers" - View it on GitHub