Discussions

I would like to know if there is any setting/config/param that might help to have deterministic behavior (same input == same output).

Knowing that some AI models exhibit non-deterministic behavior. This means that even if you provide the same prompt, the model's internal processes may produce different outputs each time it's run. This behavior is inherent to certain machine learning models and is not specific to the Perplexity AI service.

While searching on the internet, the output will depend on the available data, but at least I would like to achieve that for the same request (with exactly the same payload) I get same output.

(I was looking into the docs to see if there is any additional configuration apart of the model and messages, like seed or other parameters to ensure deterministic behavior)

Thanks in advance.