Unverified50% confidenceFactExact time
CoRD's design automatically anneals the replacement strength over training: as the student gradually learns the teacher-preferred tokens, the two distributions converge and the replacements naturally decrease.
1
Sources
50%
Confidence
Long-term
Relevance
11/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedD²-OPD determines the extrapolation direction by adaptively weighted combination of the teacher model's post-training offset and the initial student-teacher gap.73% similarUnverifiedExisting few-step distillation methods fall roughly into two categories: one relies on an additional score model to estimate the teacher distribution, and the other introduces adversarial losses to bring the student's output distribution closer to the teacher's.71% similarUnverifiedThe paper "CoRD: Coupled Rollouts Fill the Blind Spot of On-Policy Distillation" proposes the coupled rollout distillation (CoRD) method, aimed at actively pushing the distribution of training states toward the teacher side without abandoning the online loss.69% similarUnverifiedIn the RL post-training scenario, the teacher model's output distribution keeps changing as training proceeds, and the distillation supervision signal becomes 'the relative change in the teacher policy' rather than the absolute output.68% similarUnverifiedIn conventional DMD training, the critic is required to model a step-independent mixture distribution of student outputs.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1141704API
curl https://kongchang.com/api/v1/knowledge/claims/1141704MCP
get_claim(id=1141704)