Reasoning or Rhetoric? An Empirical Analysis of Moral Reasoning Explanations in Large Language Models