GReaTer: Gradients over Reasoning Makes Smaller Language Models Strong Prompt Optimizers Paper • 2412.09722 • Published 12 days ago • 5