(Security) Assertions by Large Language Models