← Quit0 / 0 correct
LLMs & AIHard1 of 239

What is the main trade-off of quantizing an LLM to 4-bit weights?