What we learned from using GPT for 500k+ classifications

Re:What we learned from using GPT for 500k+ classifications by dontupvoteme on Tuesday July 11, 2023 @04:39PM (#36686909)

Making it rank multiple attributes on a scale of 1-10 also works decent in my experience (albeit much better with 4 than 3).

[ Parent | on HN ]

Re:What we learned from using GPT for 500k+ classifications by caesil on Tuesday July 11, 2023 @06:19PM (#36688027)

Yes, agreed. I doing this as well. Works excellently for NLP classifier tasks.

Funnily enough, there is a certain propensity for it to output round numbers (50, 100, etc.) so I have to ask it not to do this and provide examples ("like 27, 63, or 4"). Now that I think about it I should probably randomize those.

[ Parent | on HN ]

Re:What we learned from using GPT for 500k+ classifications by dontupvoteme on Wednesday July 12, 2023 @03:00PM (#36699990)

Interesting, I've just been doing 1-10 (maybe i should include 0) -- Do you get the same result if you floatify the larger integers, e.g. 0.000 - 10.000?

[ Parent | on HN ]