LAD
A locality-aware distillation module in Agentic-OPD that restricts knowledge transfer to the student's failure trajectories and only to problems on which the teacher performs better. This avoids overwriting capabilities the student has already mastered and prevents capability regression.
Timeline (last 90 days)
The LAD (Locality-Aware Distillation) module restricts knowledge transfer to the student's failure trajectories, and only to problems on which the teacher performs better.
LAD (Lightweight Ahead Drafter) is a lightweight module specifically responsible for predicting multiple future tokens using early loop states; it relies only on early loop features and has very low computational cost.
All Facts (2)
The LAD (Locality-Aware Distillation) module restricts knowledge transfer to the student's failure trajectories, and only to problems on which the teacher performs better.
50%UnverifiedLAD (Lightweight Ahead Drafter) is a lightweight module specifically responsible for predicting multiple future tokens using early loop states; it relies only on early loop features and has very low computational cost.
50%