[Shitty-kitty]

That's exactly what the LLM seems to have done as well. The problem is that we want and even expect the A.I to be truthful.