Distilling the behavior of a context-rich "teacher" instance of a model back into its own base weights, so within-session learning becomes permanent — e.g. on-policy self-distillation (OPSD).
Tensions
OPSD needs no verifiable outer reward and gives denser supervision than RL, while inheriting RL's sparse-update property that guards against catastrophic forgetting. The open question is whether matching a teacher's per-token distribution captures durable skill or merely imitates surface trajectories — and how to handle the rollout after the student diverges from the teacher's path.