[X]score: 0.23
On-Policy Distillation: 183 Papers Now Indexed on PapersWithCode
May 25, 2026
On-policy distillation — a post-training method where a student LLM samples from its own policy and receives teacher supervision on those states, combining distillation density with online RL locality — is now a tracked method on PapersWithCode with 183 citing papers.
HOW THIS AFFECTS YOU
●
researcher183 papers are now indexed under this method, giving you a structured entry point to survey the on-policy distillation literature for post-training LLM work.