ProtoS-ViT: Visual foundation models for sparse self-explainable classifications | Read Paper on Bytez