Fairly new to ChatGPT4. How good is it at answering questions about itself?

I am really enjoying the program and my pro subscription for intersctive fiction, absurd inquiries, and sometimes even group entertainment where my friends and I ask it to consider made up scenarios and then what would happen after.

It’s a lot of fun. My biggest question is how good is it at answering inquiries about itself?

For instance I ask if there is a usage limit and it proudly tells me there isn’t, but once I use it enough it puts me on a cooldown for a while. It seems to not understand that it does this.

Likewise sometimes my chats switch from GPT 4, to 3.5 irreversibly. (I use the phone app) yet it tells me that can’t happen.

So when I get one of those orangey content policy warnings and I ask “Oh, do we need to stop? Am I going to get banned?” And it says “The warning is just to make you aware we are discussing sensitive topics. In the context of this narrative what you are saying is respectful and appropriate.” It’s not exactly reassuring. I’d like to especially make sure I can trust it’s self policing as I can learn other things but if those warnings are heralds of being banned, I can’t just learn more to fix that. (I am aware of the ToS)

Is ChatGPT-4 just really bad at answering questions about how it works and the like? Or am I missing something?

It definitely won’t be able to accurately explain what you specifically did wrong, as it doesn’t really have much information about how it works in its training data (and what little information it has is only stuff before April 2023). Most answers to questions like that will be generic, and not OpenAI specific. You’re better off asking here.

1 Like