The Pluralistic Moral Gap: Understanding Judgment and Value Differences between Humans and Large Language Models