Your question is Adapting VLA for Unseen Spatial Environments. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How would you modify a VLA (Vision-Language-Action) model to better handle unseen spatial environments?