and then ChatGPT showed ChatLLaMA how it should have responded by roasting itself from LLaMA's perspective?

Now that's an alpha LLM flex.