[Oxidation]

It always acknowledged it to me. Something like:

> AI: Answer that's clearly factually wrong (e.g. a function that doesn't exist or a completely wrong numerical figure).

> Me: that's not right, the function doesn't exist (etc.)

> AI: you are right, that function doesn't exist. The answer is blah

And repeat, since if it didn't get it right first time, it seems unlikely to be able to get there at all (and you'd not know when it did unless you already know the answer).

It's very like a specific person I know in a PR-marketing type job that will just glide across a noticed outright falsehood and instantly reshape what they're saying in real time and carry on as if nothing happened, leaving you wondering if you're taking crazy pills.

[Fatnino]

I tried to get it to highlight the row with today's date in Google sheets. It got it wrong, I told it it was wrong, acknowledged and tried again. After 3 wrong answers I gave up and went to Google where the answer was the first result. I went back to chatgpt and told it "the way to do this is actually {blah}" and it went "yes, to do this you do such and such..." and launched into the regular speil where it explains the code it just gave you to you.

Mansplained by a robot...

[MuffinFlavored]

I always wondered why it bothers returning the first result if it can tell it's wrong 30 seconds later if you ask it again "I think that's wrong".

Why not do make the function that outputs answers also feed itself "is this actually right/are you sure/is this not wrong"? Too expensive? Giant loop?

[gabelschlager]

It's a language model trained to return the most likely answer given a specific prompt. The goal of the answer is, to simulate a conversation and to sound realistic. ChatGPT is not grounded in reality, it does not know where its "knowledge" comes from, nor does it know whether the things it is saying are correct or wrong.

It can't tell it's wrong, it just reacts in a plausible well to you telling it it's wrong (acknowledging it and giving another explanation). Since ChatGPT is an expansion of GPT-3, those things won't be really solved since being somewhat of a knowledge base is a nice side effect, not the main goal of the model.

[MuffinFlavored]

> It can't tell it's wrong

So you are saying the order of events if:

1. you ask it a question

2. it gives you an answer (in this case, it's wrong)

3. you ask it "are you sure? i think that's wrong"

4. it answers again "you are right, i think the answer i just gave you was wrong, here is what i think is the right answer this time" (and again, it's wrong)

and repeat forever?

[yetanotherloser]

Thank you, this is a really clear way of explaining it and a careful reading would be helpful to lots of people who think this is something it isn't.

[Oxidation]

> Why not do make the function that outputs answers also feed itself "is this actually right/are you sure/is this not wrong"? Too expensive? Giant loop?

They do this, with humans. Both during training (they use supervised and reinforcement learning), and now at a much greater scale: it's what the free public access period is for and why there's a thumbs up/down button next to the output.

[akmarinov]

It can't tell that it's wrong. You tell it that it's wrong.

Then you reset your session, you ask the same question and you get the wrong answer again.