Search papers, labs, and topics across Lattice.
4
0
6
2
Pixel Linguist II not only sets new benchmarks in multilingual visual text understanding but also maintains performance under 80% visual token compression.
AtomiMed achieves a strikingly higher correlation with human radiologist evaluations, transforming how we assess clinical report accuracy.
A tailored vision-language framework boosts diagnostic accuracy on 3D CT imaging by over 5% while enhancing zero-shot reliability through innovative semantic alignment techniques.
LLMs can learn to tailor their queries for different retrievers, revealing that one-size-fits-all strategies are ineffective in retrieval-augmented generation.
