Prompt-Level Distillation: A Non-Parametric Alternative to Model Fine-Tuning for Efficient Reasoning Paper • 2602.21103 • Published 14 days ago