Read a draft of “How Sturdy Are Large Language Models’ Moral Foundations?” here