VLA-GAUGE: Systematically Evaluating Generalization and Robustness in Vision-Language-Action Models
Vision-Language-Action models (VLAs) have achieved high success rates on familiar manipulation tasks, yet it remains unclear whether such scores reflect genuine zero-shot generalization and perturbation robustness. Existing evaluation frameworks provide limited capability-level diagnosis and combinatorial perturbation...
Jun Xiao, Rui Ye, Dong-Run Cao et al.
· 0 citations