3 ms·
Keep in mind that the system always chooses randomly so there is always a possibility it commits to the wrong output. I don't know why openAi won't allow deter
by anon291 2y ago
Keep in mind that the system always chooses randomly so there is always a possibility it commits to the wrong output.
I don't know why openAi won't allow determinism but it doesn't, even with temperature set to zero
- coffeebeqn 2y agoWould picking deterministically help through? Then in some cases it’s always 100% wrong
- jaredsohn 2y agoYes, it is better if for example using it via an API to classify. Deterministic behavior makes it a lot easier to debug the prompt.
- maeil 2y agoNondeterminism provides an excuse for errors, determinism doesn't. Nondeterminism scores worse with human raters, because it makes output sound even more robotic and less human.
- roywiggins 2y agoDeterminism only helps if you always ask the question with exactly the same words. There's no guarantee a slightly rephrased version will give the same answer, so a certain amount of unpredictability is unavoidable anyway. With a deterministic LLM you might find one phrasing that always gets it right and a dozen basically indistinguishable ones that always get it wrong.
- anon291 2y agoMy program always asks the same question yes.