@AnthropicAI So far all models have been jailbroken, even anthropic models, so unless you find a solution to jailbreak, which afaik there isn’t, it’s pointless to keep applying nonsense guardrails that hurt not the bad actors, but the legitimate users…