Category: On-Policy Data

• RL Remembers More Firmly, SFT More Forgetful? Princeton Danqi Chen's Team Rewrites Post-Training Cognition