Awesome Robotics Manipulation · full_paper

Not All Features Are Created Equal: A Mechanistic Study of Vision-Language-Action Models

ICLRW 2026 · 2026-03-19 · Low-Level Learning-Based Action Modelling / Input Modelling / Evaluation of Vision Language Action Models

VLA机器人学习
新标签打开 AlphaXiv新标签打开 Paper新标签打开 Code
Not All Features Are Created Equal: A Mechanistic Study of Vision-Language-Action Models figure
AlphaXiv 中文论文页面（可滚动查看）