Could you please tell me which OpenAI API you used during the MT-Bench evaluation?

[hitszxs]

Could you please tell me which OpenAI API you used during the MT-Bench evaluation?
gpt-3.5 turbo or others?

[jiwooya1000]

Hello @hitszxs,
We used the default setting of the FastChat repo when we worked on this project, which makes GPT-4 an evaluator.

[hitszxs]

OK! Thank you for your response.

Hello @hitszxs, 你好 ， We used the default setting of the FastChat repo when we worked on this project, which makes GPT-4 an evaluator. 我们在处理此项目时使用了 FastChat 存储库的默认设置，这使得 GPT-4 成为评估器。