{
  "id": 4640,
  "url": "https://arxiv.org/abs/2605.09253v4",
  "title": "Cornerstones or Stumbling Blocks? Deciphering the Rock Tokens in On-Policy Distillation",
  "summary": "While recent work in Reinforcement Learning with Verifiable Rewards (RLVR) has shown that a small subset of critical tokens disproportionately drives reasoning gains, an analogous token-level understanding of On-Policy Distillation (OPD) remains largely unexplored. In this work, we investigate high-loss tokens, a token type that--as the most direct signal of student-teacher mismatch under OPD's per-token KL objective--should progressively diminish as training converges according to existing stud",
  "authors": "Yuxuan Jiang, Runchao Li, Shubhashis Roy Dipta, Dawei Li, Zhao Yang",
  "category": "research",
  "topics": "regulation,children-education,finance-investment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-10T01:41:43.000Z",
  "fetched_at": "2026-07-14T16:31:08.357Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/4640",
  "original_url": "https://arxiv.org/abs/2605.09253v4",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}