Susceptibility to Influence of Large Language Models