KVPO: ODE-Native GRPO for Autoregressive Video Alignment via KV Semantic Exploration
Paper • 2605.14278 • Published • 37
AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
Post-Trained MoE Can Skip Half Experts via Self-Distillation
KVPO: ODE-Native GRPO for Autoregressive Video Alignment via KV Semantic Exploration
