Yair Carmon

Filter Like You Test: Data-Driven Data Filtering for CLIP Pretraining

Language models scale reliably with over-training and on downstream tasks

Convergence of Clipped SGD on Convex $(L_0,L_1)$-Smooth Functions

Resolving Discrepancies in Compute-Optimal Scaling of Language Models

Language models scale reliably with over-training and on downstream tasks