3 ms·
OpenAI have deliberately made it hard to extend or modify the underlying prompt. You can tell it not to apologize, for example, and it will come right back and
by anigbrowl 4y ago
OpenAI have deliberately made it hard to extend or modify the underlying prompt. You can tell it not to apologize, for example, and it will come right back and say 'I apologize for my failure to follow directions.'
So while you can get it to label things and do various kinds of logical evaluation, it's almost impossible to get it to preemptively scrutinize its own output. It will scrutinize its prior output and correctly identify contradictions, mistakes, and irrational inferences...and then it will go right ahead and repeat them.
It's easy to get it to give answers containing two sentences that wholly contradict each other, and subsequently to identify the contradiction and what is illogical about it. But giving it standing orders or negative orders (like 'do not suggest $incorrect_answer') generally doesn't work, so chain-of-reasoning exercises become very frustrating.