Mirage bounds LLM KV-cache timing leaks without blocking sharing
On a 500-sample corpus and three models, token-emission delays pushed leakage near random guessing with lower TTFT than baselines.
2 stories, newest first · one card per story
On a 500-sample corpus and three models, token-emission delays pushed leakage near random guessing with lower TTFT than baselines.
The token comparison in Apache Struts2 Core 7.4.0 and below uses early-exit functions like != or memcmp, creating a timing side-channel proportional to matching bytes.
Today 1.5K readers · 2.4K views · 754 reads · all time 189K views