continuation_token

[glass raptor]

#community-help
Hi , I have been trying to use continuation_token parameter and context parameter as part of creating a completion. But I always get the values as empty irrespective of the prompt. Then gpt said these functions are deprecated? Then I had set my stream to True, but still didnt get any output in the responses . Anyone worked on it before, any idea how to resolve ?

[modern elm]

I’m not very familiar with this but this is the answer of ChatGPT, I hope that might help you.

The continuation_token and context parameters were indeed deprecated in GPT-3, so it's not recommended to use them in current versions of the API. Instead, you can use the history parameter to provide context to the model when generating completions.

The history parameter is a list of previous conversational turns, where each turn is represented as a string. To provide context to the model, you should include all previous turns in the conversation up to the current point as the history when making a completion request.

Regarding the issue with the stream parameter, keep in mind that setting stream=True means that the API will not return the full completion result in a single response, but will instead return intermediate results as they become available, which can be useful when generating very long completions. If you want to retrieve the full completion in a single response, you should set stream=False.

If you're still having trouble, it might be helpful to provide more details about your setup, including the specific API you're using and any code snippets that demonstrate the issue you're encountering.

[sick geyser]

There is no such "history" parameter for the API

The only way to provide context to the completions models is by having the context in the prompt