[-] henrique@lemmy.world 10 points 11 months ago

Made an integration with one LLM recently and had the same thought. Too many inputs and the responses are incorrect. Too few and it starts to hallucinate.

henrique

joined 1 year ago