A linear relationship between initial student-self-teacher performance gap and OPSD improvement provides a predictive law across contexts and model families.
Opsd isn’t a silver bullet for continual learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Predictive Law for On-Policy Self-Distillation From World Feedback
A linear relationship between initial student-self-teacher performance gap and OPSD improvement provides a predictive law across contexts and model families.