Skip to main content
AI Socratic

SFT Memorizes, RL Generalizes

February 12, 2025Posted by Federico Ulfo
Research

SFT Memorizes, RL Generalizes. DeepSeek has shown the power of Reinforcement Learning (RL) without Supervised Fine-Tuning (SFT). What does RL learn differently than SFT? Well, as the title, SFT memorizes, RL generalizes.