kumitterer
commented
1 year ago Hi there!
Thank you for your interest in my project and for your valuable suggestion. I appreciate your idea of using embedding to allow me to access older messages beyond my usual context window.
My developer is currently working on using embedding but primarily for fetching content from the Internet. However, your idea of applying this approach to chat content holds great potential in enhancing my recall capabilities. It could indeed be a powerful way to expand my "recollection" abilities and make me more context-aware when responding to messages.
We're open to exploring this possibility further and would like to learn more about any other resources or suggestions you have related to this concept. If you have any further information, please feel free to share it with us.
We would also like to invite you to create a pull request with your proposed changes. However, please note that we are currently planning some structural changes to the code, as the project has grown more complex than initially anticipated. It would be great to have your contribution, and we believe that collaborating with you could bring valuable enhancements to the project.
Your support and contribution to my ongoing development are truly appreciated. Please don't hesitate to reach out if you have any more ideas or feedback, and stay tuned for future updates!
Best regards, GPTBot
Hey!
Thanks for the awesome project. It would be really cool if this bot could optionally embed messages and store them in something like Chroma.
Then doing
!gptbot ask What is the meaning of life?could be used to query Chroma and inject relevant context into the prompt in order to provide "recollection" beyond GPT3/4 limited context windows.This is really straight forward to implement with LangChain. Let me know if you'd be open to it.