Are Large Language Models Good Classifiers? A Study on Edit Intent Classification in Scientific Document Revisions