Search papers, labs, and topics across Lattice.
3
0
5
Unlock compositional reasoning in 3D vision-language models with a new dataset that maps noun phrases to 3D instances, revealing improvements on both fine-grained and traditional segmentation tasks.
Existing multimodal models struggle with multi-image reasoning, but a new benchmark and inference-time attention fix exposes and alleviates these shortcomings.
Steer frozen MLLMs to reason about specific image regions at test time, without any training, by optimizing visual prompts that guide cross-modal attention.
