What do you think…
I guess ultimately something hype-related needs to be able to be converted into a vector space/for a vector database.
His example uses documentation, but you were already making that happen through normal means. Along those lines though, maybe all the forum contents could be pulled out?
I'd love to apply this somehow to Hype documents themselves, but I guess I'm not really sure how to do such a conversion into a vector space...?
Apart of creating the vector representation maybe a a service live Vectera can be used…
It works like a charm! I used it to answer a question, I answered today manually and it does an impressive job. It combines all my previous attempts (Hype ChatGPT Copilot and Documention mode), and still has loads of unused context… they just increased the context window from 8k to 128k (16x) while reducing costs by 2.75x…
Hi Max, is your assistant for public? Or do i have to set it up myself?
Which documents did you upload - from where are they?
Thx for info, best jan
Hello! The assistant you're seeing is part of my OpenAI playground. Presently, the ability to publish GPT models is exclusive to a limited group of ChatGPT Plus users. It’s expected that this feature will become more widely available soon. Note that Tumult holds the copyright for the essential resources needed to release this Hype Assistant model publicly. It’s likely they will launch an official GPT version, which should be quick and straightforward for them to set up.
P.S. If you have access to the playground or wish to deploy a private GPT model, refer to the documents from Hype ChatGPT and Document Mode that I've shared previously. The official Tumult Hype Documentation PDF is also a good resource. I’ve tailored the documents to remove some content to fit the previous interface limits, so feel free to adjust them according to your needs.
You are welcome to use any of our website/documentation for this provided:
Our value is in the product we produce - therefore it is in the best interest that documentation and ways of accessing this documentation are as freely available as possible. (In other words: We don't generally care about copyright on the documentation or runtime, just the proprietary source code for the app itself).