Small Language Models Improve Giants by Rewriting Their Outputs Paper • 2305.13514 • Published May 22, 2023 • 2