LeftoverLocals: Listening to LLM Responses Through Leaked GPU Local Memory.

Tyler Sorensen Heidy Khlaaf

Published in: CoRR (2024)

Keyphrases

memory usage
real time
memory requirements
graphics processors
memory bandwidth
graphics hardware
multi threaded
low memory
data sets
search engine
main memory
parallel implementation
parallel computing
intel xeon
memory size
memory space
data transfer