Order-Only On-Policy Distillation
READ THE ORIGINAL SOURCEOpen original sourcehuggingface.co · Opens in a new tabExternal source links are regularly re-checked for safety, and checked again before opening using a recent cached result when available. How link checks work →
The paper proposes Order-Only On-Policy Distillation (O³PD), which trains a student at each generated prefix to reproduce the teacher's ordering of its top-K tokens rather than matching token probabilities. The objective is invariant to teacher temperature and needs no temperature tuning. Across six teacher-student pairs evaluated on HMMT25, AIME24, and AIME25, O³PD outperforms standard OPD on average in every pair, by up to 2.9 points, with gains up to 5.6 points on AIME25. Other ranking losses give similar gains.
One primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face Daily Papers
- Discovery linkHugging Face
- Linked in sourceView arXiv page
- Linked in sourceView PDF
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Open full article view →- Original paper+45
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
The paper proposes Order-Only On-Policy Distillation (O³PD), which trains a student at each generated prefix to reproduce the teacher's ordering of its top-K tokens rather than matching token probabilities. The objective is invariant to teacher temperature and needs no temperature tuning. Across six teacher-student pairs evaluated on HMMT25, AIME24, and AIME25, O³PD outperforms standard OPD on average in every pair, by up to 2.9 points, with gains up to 5.6 points on AIME25. Other ranking losses give similar gains.
A READER'S EVIDENCE CHECKHow to read this signalSix questions about claims, sources, risks, and open checks
This is a signal-level guide until claim-level evidence is available. “Not recorded” means the evidence has not been assessed here; it does not mean a result is false. How we check AI news →
- 01Who is making each claim?
The claimant and whether each claim was measured or asserted are not recorded in this public brief.
- 02What was actually tested?
Not recorded per claim. Check the source for sample size, tasks, model and benchmark versions, and reproducibility.
- 03Where are the sources?
Original link: huggingface.co. Brief basis: fetched source page. Primary page date and quote locations are not established in this panel.
- 04What could mislead?
Source claims not independently checked; results are reported by the authors on specific math reasoning benchmarks.
- 05What is at risk if wrong?
Not assessed per claim. Consider safety, privacy, cost, and adoption decisions before relying on the result.
- 06What would change the view?
No claim-specific check is recorded. An inspectable primary method or independent replication could change the assessment.
Key findings from the stored brief
- O³PD trains the student to reproduce the teacher's order of its top-K tokens without fitting probability values.
- Across six teacher-student pairs, O³PD beat OPD on the HMMT25/AIME24/AIME25 average in every pair, by up to 2.9 points.
- Gains reached 5.6 points on AIME25, and other ranking losses yielded similar improvements.
- The authors report O³PD departs further from teacher probabilities yet reproduces the teacher's ranking more faithfully.
Source claims not independently checked; results are reported by the authors on specific math reasoning benchmarks.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
A publication revision history is not stored for this signal. New engagement snapshots do not mean the source itself changed.
Why this may matter
It suggests that token ranking, rather than calibrated probability values, may carry much of the knowledge transferred in on-policy distillation, and removes the need to tune teacher temperature for this objective.
Limits and open questions
Source claims not independently checked; results are reported by the authors on specific math reasoning benchmarks.
Organization-specific financial, security, and operational effects require the organization's own context and review.
Join the discussion
Have a different perspective or a question about the evidence? Discussion is free with a verified email account. Every comment is reviewed before publication.
No published comments yet.
New commenters: 2 submissions per 24 hours, at least 5 minutes apart, until 3 are approved. Discussion guidelines and moderation →