from Hacker News

GPT-4 Trained on 10x More Data, Can Write 60k Word Books

by baptiste313 on 1/11/23, 6:42 PM with 1 comments

  • by minimaxir on 1/11/23, 6:56 PM

    This misinformed GPT-4 take is new levels of comically bad because it correlates number of model hyperparameters with the quantity of training data.