Search papers, labs, and topics across Lattice.
1
0
3
Current multimodal agents still struggle to combine ambiguous visual cues with open-web verification, highlighting a critical gap in their ability to perform complex geolocation tasks.
