题意："什么时候我可以在 OpenAI 的回复中预期到多个“选择”？"

问题背景：

For a simple OpenAI api call in python

对于一个在 python 中简单的 OpenAI API 调用

response = openai.ChatCompletion.create(
    model="gpt-4",
    messages=[
        {"role": "system", "content": instructions},
        {"role": "user", "content": my_input}
    ])

I get meaningful data in response['choices'][0]['message']['content']

我在 response['choices'][0]['message']['content'] 中获取到有意义的数据

Under what circumstances will there be multiple entries in response['choices']? How should I interpret this?

在什么情况下 response['choices'] 中会有多个条目？我应该如何理解这一点？

问题解决：

From what I see, the documentation says that if you set the parameter n to more than 1 you will get more than one "choice".

"据我所知，文档中提到，如果将参数 n 设置为大于 1，你将会获得多个 '选择'。

https://platform.openai.com/docs/api-reference/chat/create

n integer or null Optional Defaults to 1

How many chat completion choices to generate for each input message. Note that you will be charged based on the number of generated tokens across all of the choices. Keep n as 1 to minimize costs.

I don't know the concept behind naming it choices, but I can tell you that the default is 1, so you always will have to access index 0 from choices unless you change the n parameter.

"我不清楚将其命名为 choices 背后的概念，但我可以告诉你，默认值是 1，所以你总是需要从 choices 中访问索引 0，除非你更改了 n 参数。"