SECURING LARGE LANGUAGE MODELS AGAINST JAILBREAKING ATTACKS: A NOVEL FRAMEWORK FOR ROBUST AI SAFETY. LEX. 2025;23(S6):558-567. doi:10.52152/801808