Today I built an app I had wanted to build for a very long time. It's basically a personality-based AI chatbot. You provide 2 simple things: 1st, instructions on how the model should behave. You can be very specific. 2nd, data. You want the personality to understand, that data can be bunch of PDFs. It then turns that data & instructions into a personality. So, next time you ask it anything, that personality will have all the context of the data you provided & will behave according to your instructions. When I'm all set, I'll open source it. you can simply attach your own API. image

Replies (11)

alenasatoshi's avatar
alenasatoshi 10 months ago
How much text is 2mil tokens? Ie If I need the personality to read about 5000 pages of literature before we can get along, is that possible?
2 million context token should be enough for 5,000 pages. It should need around 1.5 million tokens, leaving a lot of extra context window. Then you will basically save that model state, my app turns that into personality. After that, you can use that personality as often as needed.
I'm currently using Gemini models with a mighty 2 million token context window. I use a portion of this context as a knowledge base, then employ a system prompt to alter the model's behavior accordingly. I've built many rage apps before, but with this one, I wanted to try something new: something easily self-hosted at no cost, significantly lightweight, & storing all personality & chat history locally. For more details, read this note1fk0w6a4sqz2gtsphg42z54l546d368z0700gmp8yu0tgp6ng7dmqz4fku8
Thank you. Very interesting. I'd like to play with something like this myself. How do I start? Get Gemini pro and upload my instructions as a long system prompt? Is it similar to custom GPT?
↑