Re:Show HN: Explore large language models on any computer wi...
by jncraton on Saturday June 17, 2023 @06:43PM
(#36375488)
Thanks for pointing that out. Classification is half-baked at the moment. It should ultimately be restricting output to only appropriate labels, but right now it is simply sampling.
Though, surprising results from repl.it: lm.classify("unabridged", "positive", "negative")=="unabridged is"