On-Policy Distillation with Negative-Policy Rollouts
NP-OPD improves on-policy distillation by using weaker-model rollouts as an explicit negative signal.
1 stories, newest first · one card per story
NP-OPD improves on-policy distillation by using weaker-model rollouts as an explicit negative signal.
Today 251 readers · 270 views · 20 reads · all time 108K views