Enhancing Contextual Understanding in Large Language Models through Contrastive Decoding