Towards Understanding How Knowledge Evolves in Large Vision-Language Models | Read Paper on Bytez