undefined | Better HN

0 pointsnprateem2y ago0 comments

I was trying to create an AWS IAM policy restricted to assumed roles. I didn't know you can't use assumed roles in principalArn conditions blocks but must refer to the IAM role instead. Gpt4 happily wasted an hour by shuffling the conditions around etc instead of telling me this. Sometimes it's policies were even malformed, and in all cases they didn't work.

0 comments

ilaksh2y ago

Did you try giving GPT4 the relevant documentation akong with your query?

nprateemOP2y ago

Without wanting to sound arsey, why should I have to? If I'd had the relevant docs to hand I wouldn't have needed it.

This is the problem IMO. The model needs to somehow learn that out of its entire training set, the single sentence in the AWS docs saying not to use the assumed role ARN takes precedence over any patterns it may have learnt elsewhere in this specific situation.

yodon2y ago

What you are describing is very different from the hallucination behavior of GPT-3 and GPT-3.5.

Yes, GPT-4 came up with an incorrect answer, but it's an incorrect answer an experienced programmer could legitimately have come up with, and one they probably would have come up with before actually testing their code against the AWS endpoints. GPT-4 sometimes gets hard questions wrong. GPT-3 and GPT-3.5 make up nonsense.

If a coworker told you GPT-4's answer, you'd say they were wrong but you wouldn't say they were hallucinating. If a co-worker gave you GPT-3 or GPT-3.5's answer you'd definitely doubt their sanity.

1 more reply

j / k navigate · click thread line to collapse