Past Performance is Not Indicative of Future Results (2020)

Re:Past Performance is Not Indicative of Future Results (2020) by lumost on Saturday July 31, 2021 @09:08PM (#28022851)

The quirky thing to remember about gpt-3 is that it really is just a giant autocomplete based on the internet. It can do math insofar as it’s memorized some text which did that math with slightly different verbiage etc.

If you ask it to compute something that would never have been seen on the internet it’s likely to fail. E.g. add 2 extremely large/rare numbers together

[ Parent | on HN ]

Re:Past Performance is Not Indicative of Future Results (2020) by vladTheInhaler on Monday August 02, 2021 @10:49AM (#28037208)

As another poster indicates, that's specifically not the case. Most possible arithmetic problems of reasonable size aren't anywhere in the dataset, but it can solve them with pretty good accuracy.

[ Parent | on HN ]

Re:Past Performance is Not Indicative of Future Results (2020) by fossuser on Sunday August 01, 2021 @09:22PM (#28031940)

You should read the excerpts in the paper I link to which suggest otherwise (that it’s not memorized and that it’s deducing rules).

[ Parent | on HN ]