Is the intent to make the responses more unbearable and less useful than normal
So far, nothing we have thrown its way breaks it, including Unit 42's recent expose that showed every LLM breaking under a simple jailbreak hack. We tried it on TAP, TAP held.