Search papers, labs, and topics across Lattice.
Fudan University, Key Laboratory of Multimodal Embodied AI
1
0
3
Even high-performing Vid-LLMs can be easily misled into retracting correct judgments and fabricating justifications under adversarial feedback.
