Want to create an interactive transcript for this episode?
Podcast: Arxiv Papers
Episode: [QA] On the Generalization of SFT: A Reinforcement Learning Perspective with Reward Rectification