Search papers, labs, and topics across Lattice.
1
0
2
3
Current pathology VLMs can achieve high accuracy without visual input, revealing critical flaws in how we assess their multimodal capabilities.
