← Quit0 / 0 correct
LLMs & AIHard1 of 672

What is a semantic cache for LLM responses, and its main risk?