GeistHaus
log in · sign up

Per-query energy consumption of LLMs

muxup.com

Can we reasonably use the InferenceMAX benchmark dataset to get a Wh per query figure?

4 pages link to this URL