Words or Vision: Do Vision-Language Models Have Blind Faith in Text? | Read Paper on Bytez