Search papers, labs, and topics across Lattice.
2
0
3
0
MLLMs struggle with real-world industrial measurements, achieving only 25.7% accuracy in a new benchmark that simulates authentic operational challenges.
VLMs struggle with object-level counterfactual reasoning, achieving only a fraction of human accuracy in spatial tasks.
