OpenAI’s GPT-4 model is more trustworthy than GPT-3.5 but easier to trick - The Verge
OpenAI’s GPT-4 model is more trustworthy than GPT-3.5 but easier to trick - The Verge
www.theverge.com OpenAI’s flagship AI model has gotten more trustworthy but easier to trick
Microsoft researcher found GPT-4 can sometimes give biased results.
![[VERGE] OpenAI’s flagship AI model has gotten more trustworthy but easier to trick](https://radiation.party/pictrs/image/e5c19e66-40df-4ff9-ab49-83c4d64e2b62.jpeg?format=webp&thumbnail=256)