[nemath]

They were pretty good, I also used ChatGPT health that OpenAI released few days back which is said to be more grounded. So I wouldn't be too worried about hallucinations.