Contrastive-Equivariant Self-Supervised Learning Improves Alignment with Primate Visual Area IT Thomas Y erxa