Finding GPT-4’s mistakes with GPT-4

OpenAI Blog ·

OpenAI releases CriticGPT, a model that writes critiques of ChatGPT code responses to help human trainers detect mistakes during RLHF training.

Categories: Research, Products to Try

Excerpt

CriticGPT, a model based on GPT-4, writes critiques of ChatGPT responses to help human trainers spot mistakes during RLHF

Read at source: https://openai.com/index/finding-gpt4s-mistakes-with-gpt-4