{
  "id": 4752,
  "url": "https://arxiv.org/abs/2605.07396v1",
  "title": "Rubric-based On-policy Distillation",
  "summary": "On-policy distillation (OPD) is a powerful paradigm for model alignment, yet its reliance on teacher logits restricts its application to white-box scenarios. We contend that structured semantic rubrics can serve as a scalable alternative to teacher logits, enabling OPD using only teacher-generated responses. To prove it, we introduce ROPD, a simple yet foundational framework for rubric-based OPD. Specifically, ROPD induces prompt-specific rubrics from teacher-student contrasts, and then utilizes",
  "authors": "Junfeng Fang, Zhepei Hong, Mao Zheng, Mingyang Song, Gengsheng Li, Houcheng Jiang et al.",
  "category": "research",
  "topics": "regulation,safety-alignment,children-education",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-08T07:52:15.000Z",
  "fetched_at": "2026-07-14T16:31:12.747Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/4752",
  "original_url": "https://arxiv.org/abs/2605.07396v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}