@norvid_studies (a slightly unfair gloss of his post, but if you haven't looked at the pretraining datasets, you do not have an intuition for just how narrow of a slice llms can zero in on and…

Posters: voooooogel. voooooogel

2 likes recorded by the source corpus.