Suprmind reduces AI hallucinations by running five leading models—GPT-4,...

https://www.protopage.com/donna-zhou23#Bookmarks

Suprmind reduces AI hallucinations by running five leading models—GPT-4, Claude, PaLM 2, LLaMA 2, and Falcon—together in one shared thread. It uses an independent verifier to cross-check outputs, achieving up to 99.1% accuracy