Evaluating the Quality of Code Comments Generated by Large Language Models for Novice Programmers