Can Large Language Models Identify Authorship