Taxonomy-Aware Evaluation of Vision-Language Models | Read Paper on Bytez