Re:HellaSwag: 36% of this popular large language model bench...
by PraetorianGourd on Monday December 05, 2022 @10:27PM
(#33875527)
Agreed, but I am concerned over the chance that lots of these models are used in critical scenarios without much validation of the data set.
I'm guessing it's just Mechanical Turk content which wasn't even spellchecked.