Stabilizing On-Policy Distillation for MLLM Reasoning with Global Normalization

Add code
Jun 08, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: