I don't dismiss the transformer architecture nor auto-regressive models

I dismiss the claims that these models are approaching biological levels of intelligence & that scaling up an auto-regressive transformer would lead to human level intelligence

Comments