I have integrated ChatGPT with a cloud application (ServiceNow) and also fine-tuned davinci model but I had a general query if there is any provision in ChatGPT where it can pull data on it’s own from an application and transform it to JSON that it understands. I have a personal account so I know that this option is not there but hoping if someone has a corporate account then they can clarify if this data pulling feature exists in corporate accounts.
ChatGPT is a LLM trained on data with a cutoff date in 2021. Meaning it has no knowledge of events after its cutoff date. It also currently does not have access to internet.
None of the GPT-3 models can pull data on their own from an application unless there’s an interface that allows fetching data and sending it to the models, and there’s code that fetches data and sends to the model, and the model has been prompted/finetuned on the structure and nature of data.
It’s a better idea to process JSON data before sending and send only the required data to the API. Saves cost, time and compute.
I’m pretty sure that there’s no such feature on corporate org accounts, but still OpenAI Support should be able to help answer your question,