Unverified50% confidenceFactExact time
LEAF-OPD adopts a teacher-student architecture in which a frozen teacher model provides token-level supervision on responses generated by the student itself, following the on-policy distillation approach.
1
Sources
50%
Confidence
Long-term
Relevance
11/17/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedVersion 2.0 switched to online policy distillation, where the model generates its own responses and a teacher model scores each token; two teachers were set up — the original model plus hidden human instructions for chat, and a pure base model for instruction following, tool calling, and coding.79% similarUnverifiedThe PIVOT method inserts online corrections from a teacher model into student trajectories and applies token-level KL supervision to the correction segments.73% similarUnverifiedAsynchronous OPD means that student sampling and teacher scoring/distillation updates are not executed on the same synchronous pipeline, but are advanced in parallel by different processes or different environment instances and then exchange data through buffers or parameter servers.70% similarUnverifiedLoopOPD is a cross-loop on-policy distillation framework that uses a frozen terminal-loop policy as a computationally privileged teacher to guide the student policy at intermediate loops.69% similarUnverifiedABACUS uses a generative teacher to design judgment patterns and composition programs offline, while the judgment model solves local semantic problems online, and the saved programs compose decisions into the final answer.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1179598API
curl https://kongchang.com/api/v1/knowledge/claims/1179598MCP
get_claim(id=1179598)