Post

currently writing an eval for how to make LLMs Not Talk Like That this is actually for a hackathon in a week, but i don't have to worry about anyone swiping this idea because writing this eval is so fucking hard no one else would willing spend this much time on it lmao

i think i'm up to like 6 hours of attended work and god knows how many judge runs the hard part isn't the corpus, it's the judges!

ok thinking about it more it might be closer to 10 or 15 ...