Hugging Face's logo
Activity Feed

AI & ML interests

None defined yet.

Papers

Open-MOPD: Diagnosing and Fixing Capability Imbalance in Multi-Teacher On-Policy Distillation

Weak-to-Strong Generalization via Direct On-Policy Distillation

View all Papers

Organization Card
Community About org cards

Edit this README.md markdown file to author your organization card.

Collections 6

Open-MOPD: Multi-Teacher Online Policy Distillation
Open-MOPD's 3B family: MixSFT initialization, math/code/IF RL teachers, final multi-teacher distilled student, and training/evaluation data.
Open-MOPD: Multi-Teacher On-Policy Distillation
Open-MOPD 3B family: MixSFT, math/code/IF RL teachers, the final on-policy-distilled student, and training/evaluation data.
Open-MOPD: Multi-Teacher Online Policy Distillation
Open-MOPD's 3B family: MixSFT initialization, math/code/IF RL teachers, final multi-teacher distilled student, and training/evaluation data.
Open-MOPD: Multi-Teacher On-Policy Distillation
Open-MOPD 3B family: MixSFT, math/code/IF RL teachers, the final on-policy-distilled student, and training/evaluation data.
View 6 collections