Found 1 document(s)
Page 1 of 1
-
Vision-language-action (VLA) models have achieved strong performance on standard robotic manipulation benchmarks. However, their reliability can degrade substantially when task conditions deviate from the nominal setting...