Does On-Policy Distillation Really Distill? From Noisy Teacher to Self-Improvement Paper • 2608.31046 • Published 6 days ago • 140
Open-MOPD: Diagnosing and Fixing Capability Imbalance in Multi-Teacher On-Policy Distillation Paper • 2608.19098 • Published 18 days ago • 23
Open-MOPD: Multi-Teacher On-Policy Distillation Collection Open-MOPD 3B family: MixSFT, math/code/IF RL teachers, the final on-policy-distilled student, and training/evaluation data. • 6 items • Updated 19 days ago • 5