Attention-based Interpretability with Concept Transformers.

Mattia Rigotti, Christoph Miksovic, Ioana Giurgiu, Thomas Gschwind, Paolo Scotton

Browse the full ICLR paper archive.