Blog
Robótica & RL
SMOPD: Selective Token-Entropy Masking for Dirty-History Multi-Turn On-Policy Self-Distillation
arXiv:2608.14647v1 Announce Type: new Abstract: Dirty-history rollouts make multi-turn on-policy self-distillation (OPSD) brittle: once a student emits an erroneous intermediate reply, later turns are conditioned on that reply, and uniform distillation can spend loss on tokens that carry little corrective signal. We introduce SMOPD (Selective Masking for On-Policy Distillation), a loss-only stabilization method for multi-turn OPSD. For each generated middle-turn reply, SMOPD ranks token position...
arXiv cs.LG
·Chenyang Jiang, Changhan Huang
·
// relacionados
Leia também
Blog
Flight attendants freaked out that Google is buying tons of Spirit employee data
Blog
Attackers are using AI to build exploits for industrial control systems, U.S. agencies warn
Blog
AI labs are failing to keep their own systems in check
Editorial