Adaptable Moral Stances of Large Language Models on Sexist Content: Implications for Society and Gender Discourse