4/19/2026 at 11:50:30 AM
Interesting Approach. Curious about the latency tradeoff: OLS + SVD are much heavier than Top-K.Have you benchmarked end-to-end inference latency?by vivahir215
4/19/2026 at 11:57:37 AM
[dead]by jchandra
4/19/2026 at 11:35:23 AM
High-Fidelity KV Cache Summarization Using Entropy and Low-Rank Reconstruction
https://jchandra.com/posts/hae-ols/by jchandra