Assessing and Learning Alignment of Unimodal Vision and Language Models | Read Paper on Bytez