Causal Attention for Vision-Language Tasks | Read Paper on Bytez