Generate saliency maps from images
Is there an updated version of this blogpost? On the state of MoEs?