pith. sign in

Self-distilled reasoner: On-policy self-distillation for large language models

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

citation-role summary

background 1

citation-polarity summary

fields

cs.AI 1

years

2026 1

verdicts

UNVERDICTED 1

roles

background 1

polarities

background 1

representative citing papers

The Many Faces of On-Policy Distillation: Pitfalls, Mechanisms, and Fixes

cs.AI · 2026-05-11 · unverdicted · novelty 5.0

On-policy self-distillation fails for instance-specific privileged information because the student learns an aggregated PI-free policy, while on-policy distillation is sensitive to teacher choice and loss formulation, with stop-gradient and stabilized methods as mitigations.

citing papers explorer

Showing 1 of 1 citing paper.

  • The Many Faces of On-Policy Distillation: Pitfalls, Mechanisms, and Fixes cs.AI · 2026-05-11 · unverdicted · none · ref 7

    On-policy self-distillation fails for instance-specific privileged information because the student learns an aggregated PI-free policy, while on-policy distillation is sensitive to teacher choice and loss formulation, with stop-gradient and stabilized methods as mitigations.