1 paper
Yang Li, Youssef Emad, Karthik Padthe +8
Recent work has shown that distilling reasoning traces from a larger teacher model via supervised finetuning outperforms reinforcement learning with the smaller student model alone…