04:00
2026-07-23
arxiv.org
large-language-models
Spectral-LSH: Sub-Quadratic Prompt Compression via Krylov-Projected Locality-Sensitive Hashing
Researchers propose Spectral-LSH, a training-free prompt compression method that uses Krylov subspace methods and SimHash to reduce long-prompt inference costs by grouping similar tokens into macro-toβ¦