Hacker News new | ask | show | jobs
by dicefordeath 1147 days ago
"Vicuna is a fine-tuned version of LLaMA that matches GPT-4 performance."

Oh is that right?