Restoring Linguistic Grounding in VLA Models via Train-Free Attention Recalibration
返回详情
VLA / Vision-Language-Action 每日论文卡