Preprint2026
Weak-to-Strong On-Policy Distillation
Fangxu Yu, Zinan Lin, Xiaodong Liu, et al.
Weak-to-Strong On-Policy Distillation (W2S-OPD) improves a strong student LLM by distilling from a proxy teacher built from contrast pairs of weaker models, enabling capability transfer without a larger teacher.
0Jul 28, 2026ReasoningBenchmarks
arXiv