I have experienced this too… I have attempted prompts that have responses that are simply untrue, but also remember it is only capable of using the algorithm it’s given, so if I cannot provide a prompt that is sufficient, I cannot expect it to respond sufficiently.

Oh, and it won’t remember a name it is given outside of that prompt, which I find hilarious that each day I go through a naming session. I always stick with Zephr, while it gives me like 20 other names. 😂