Use case guidelines - Summarization token limitations

In the use case requirements it states that “Summarizers that end-users can submit any content they wish to are generally not approved. We sometimes approve summarizers with a maximum of a paragraph input (150 tokens)…”

Say I am summarising notes from a university lecture that extend perhaps 300-600 tokens. As stated this would not be approved as technically the student could submit any content.

However, is there a way, other than the content filter, to apply layered prompts to keep the summary within an acceptable scope that would be approved? If so, please point me in the right direction.

1 Like

Hi Dean, that’s a great question! I would check out the Classification endpoint or try using classification prompts (via Completions) to classify user input as lecture-related or non-lecture-related.


Beautiful! I’ll get to designing classifications prior to running prompts.

Thanks for the suggestion @joey


It was more so an example. Creating study material from a students notes is more what I was going for.

1 Like