Corentin Delacour

Uni- versity of California

Papers on Lattice

Total citations

Topics

Publication activitypapers/week, last 8 weeks

Research focus

Architecture Design (Transformers, SSMs, MoE) (1)Inference & Quantization (1)Training Efficiency & Optimization (1)

Frequent co-authors

Kyle Lee (1)Kevin Callahan-Coray (1)Kyle Jiang (1)Can Yaras (1)

Papers (1)

May 3, 2026

Uni- versity of California2w ago·also CMU ML, UMich

Stochastic Sparse Attention for Memory-Bound Inference

Attention bottlenecks in long-context decoding? SANTA slashes memory bandwidth demands by stochastically sampling value vectors, achieving 1.5x speedups without sacrificing accuracy.

Kyle Lee, Corentin Delacour, Kevin Callahan-Coray +5

Architecture Design (Transformers, SSMs, MoE)Inference & Quantization Training Efficiency & Optimization

Search

Corentin Delacour

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (1)