Towards aligned body representations in vision models

Open in new window