Video LMMs name objects and actions reliably but fail to detect the precise frames and locations of contact and release events, revealing shortcut learning instead of physical grounding.
Llm-enhanced action-aware multi-modal prompt tuning for image-text matching.arXiv preprint arXiv:2506.23502, 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Action Without Interaction: Probing the Physical Foundations of Video LMMs via Contact-Release Detection
Video LMMs name objects and actions reliably but fail to detect the precise frames and locations of contact and release events, revealing shortcut learning instead of physical grounding.