Self-Supervised Vision Transformers Learn Visual Concepts in Histopathology