DemoPSD: Disagreement-Modulated Policy Self-Distillation
On-policy self-distillation (OPSD) has emerged as a practical method for training large language models (LLMs) to reason, where a single model acts as both the teacher and the student with different levels of information access. However, recent studies have found that the teacher's dense token-level supervision, conditioned on privileged information, can lead to overfitting to in-domain patterns, suppress exploration, and hurt cross-domain generalization, while also introducing a more fundamental issue: *privileged information leakage*, where the student encodes answer-dependent shortcuts that
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 62%chrisliu298/awesome-on-policy-distillation →
- PossiblePossibly related (embedding) · 62%nick7nlp/Awesome-LLM-On-Policy-Distillation →
- PossiblePossibly related (embedding) · 54%chrisliu298/awesome-llm-unlearning →
- PossiblePossibly related (embedding) · 50%IEEE Rolls Out Large Language Models Virtual Training Course →
- LinkedLinked via arxiv author · 85%Yunhe Li →
“DemoPSD: Disagreement-Modulated Policy Self-Distillation”
- LinkedLinked via arxiv author · 85%Hao Shi →
“DemoPSD: Disagreement-Modulated Policy Self-Distillation”
- LinkedLinked via arxiv author · 85%Wenhao Liu →
“DemoPSD: Disagreement-Modulated Policy Self-Distillation”
- LinkedLinked via arxiv author · 85%Mengzhe Ruan →
“DemoPSD: Disagreement-Modulated Policy Self-Distillation”