Trust-Region Behavior Blending for On-Policy Distillation 事件

Trust-Region Behavior Blending for On-Policy Distillation arXiv:2605.31159v1 Announce Type: cross Abstract: On-policy distillation (OPD) trains a student on prefixes sampled from its own policy while matching a stronger teacher. This addresses the prefix mismatch of offline distillation, but early student rollouts can still be poor, placing teacher supervision on weak or low-quality prefixes. We propose Trust-Region behavior Blending (TRB), a warmup method that replaces the early rollout policy