I came across this new finetuned model based on Openchat 3.5 which is apparently trained used Reinforcement Learning from AI Feedback (RLAIF).
https://huggingface.co/berkeley-nest/Starling-LM-7B-alpha
Check out this tweet: https://twitter.com/bindureddy/status/1729253715549602071
Every other model nowadays claims to be GPT-4, and they turn out to be < GPT-3. I don’t know what kind of test they use to score .
LOL GPT4
https://preview.redd.it/fy2rvgg8v13c1.png?width=1754&format=png&auto=webp&s=8df41b305a0d01be335f406a204b1061ca24b658