Research on vulnerabilities in LLM serving systems that can be exploited through timing attacks and other side-channel techniques to extract sensitive information

This presentation covers 4 research papers on large language models applied to Side-Channel Attacks in LLM Infrastructure.