How Easily do Irrelevant Inputs Skew the Responses of Large Language Models? | Read Paper on Bytez