We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
The evaluation component includes Bert Score, GPT-4 ranking, and GPT-4 preference win rate.