Search papers, labs, and topics across Lattice.
3
0
8
3
Diffusion models can run over 3x faster without retraining or architectural changes, thanks to a clever second-order prediction scheme.
Cut sparse attention indexing costs by 75% without sacrificing quality by intelligently reusing indices across layers.
LLMs struggle to extract and reason over complex text structures in scientific domains, but explicitly prompting them to build these structures first yields significant performance gains.
