AI.info

Jincheng Yu

Explore Jincheng Yu on AI.info.

• MHLA: Restoring Expressivity of Linear Attention via Token-Level Multi-Head