# Mechanistic Study of VLA Models: Not All Features Are Created Equal

- Source: HuggingFace Daily Papers
- Published: 2026-10-20T00:00:00.000Z
- RoboHot: https://robohot.news/en/items/haabureyjs7732hxtzk9wshl5
- Original: https://huggingface.co/papers/2603.19233

## Summary

The study proposes an interpretability pipeline and finds that injecting task A's activations into the action expert causes pi0.5 to execute task A's motor trajectory in 99.6% of episodes (n=1,968), and that the injected program is bound to absolute workspace coordinates while ignoring the visible scene, offering a mechanistic explanation for the perturbation brittleness reported by benchmarks.