Wiring Matters: Injection Topology and Initialization of Affordance Heads in Vision-Language-Action Policies

Excerpt

arXiv:2610.06318v1 Announce Type: cross Abstract: Dense affordance supervision is an appealing auxiliary signal for vision-language-action (VLA) policies, yet naively co-training an affordance head can severely damage instruction following. We present a controlled study of how to wire such a head into a modern VLA on the LIBERO benchmark. Our recipe reads the backbone through a stop-gradient and re-injects an intermediate head feature into the action expert via a learned bridge. The stop-gradien