@_wallfacer@damn_then@Polymarket Do you see the problem of benchmarking with textbook problems? Language models are literally trained on textbooks. As soon as you go a bit outside the usual types of questions its performance drops drastically in my experience.
@shydolph@Lord_Tomas1@marievol6 Not really except some parts of northern Norway. This is largely a myth based on a misunderstanding of what indigenous means.