Data Mixing Made Efficient: A Bivariate Scaling Law for Language Model Pretraining Paper • 2405.14908 • Published May 23 • 11
