논문해외
Every Token Needs Its Own Teacher: Multi-teacher Adaptive On-policy Self-distillation
HF Daily Papers
On-policy Self-distillation (OPSD) has proven effective for post-training large language models by constructing a teacher conditioned on privileged references.
원문의 저작권은 HF Daily Papers에 있습니다. 위 발췌는 원문 확인을 돕기 위한 짧은 인용입니다.