LIBERO-PRO shows VLA models collapse from over 90% to 0% accuracy under perturbations in objects, states, instructions, and environments, exposing memorization instead of genuine comprehension.
Exploring the adversarial vulnerabilities of vision- language-action models in robotics, 2025a
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
LIBERO-PRO: Towards Robust and Fair Evaluation of Vision-Language-Action Models Beyond Memorization
LIBERO-PRO shows VLA models collapse from over 90% to 0% accuracy under perturbations in objects, states, instructions, and environments, exposing memorization instead of genuine comprehension.