Search papers, labs, and topics across Lattice.
University of Waterloo
3
0
6
A single checkpoint can now adapt to any model size, streamlining the deployment of elastic retrieval systems and achieving faster performance without sacrificing quality.
Achieving high-quality reranking with a 30B MoE model is now feasible on an academic budget, outperforming traditional dense models in efficiency.
MPE achieves superior long-context retrieval by efficiently encoding document chunks while maintaining critical contextual relationships, outperforming traditional methods.
