Unverified50% confidenceFactExact time
For correct rollouts, APSD adopts an interpolation strategy, with the target distribution lying between the student and the self-teacher, preserving the student's already-successful generation behaviors.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn-policy distillation requires the student to first generate its own prefix trajectory, after which the teacher provides guidance at the states the student actually reaches, so the state distributions are aligned, and it is usually more stable on long-horizon tasks where errors tend to accumulate.72% similarUnverifiedThe key mechanism of RM-OPD is applying token-level dense supervision on critiques generated by the student itself, making the supervision distribution consistent with the student's current policy distribution and alleviating the training-inference distribution mismatch problem.71% similarUnverifiedSelf-Distillation Policy Optimization (SDPO) provides dense token-level supervision signals to the student model through a feedback-conditioned "self-teacher".71% similarUnverifiedAPSD's experimental results show that it outperforms reinforcement learning baselines and self-distillation baselines in overall performance.69% similarVerifiedOnline policy distillation (OPD) has the student model generate its own training responses, which are then supervised by the teacher model.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1074085API
curl https://kongchang.com/api/v1/knowledge/claims/1074085MCP
get_claim(id=1074085)