Do Vision-Language Models Really Understand Visual Language? | Read Paper on Bytez