TopoNets: High performing vision and language models with brain-like topographyarxiv.org225 pointsmayukhdeba year ago