TY  - RPRT
TI  - Do Vision-Language-Action Models Understand Instructions? A Mechanistic Interpretability Study on Language Grounding
AU  - Theodor Wulff
AU  - Angelo Cangelosi
PY  - 2026
UR  - https://arxiv.org/abs/2610.10178
ID  - 2610.10178
ER  -