Search papers, labs, and topics across Lattice.
1
0
2
7
Multimodal models struggle with consistency, showing significant judgment discrepancies between text and speech inputs, especially in Arabic contexts.
