Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Update retrieval description prompt to use retrieval with newer models. #346

Open
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

ptgoetz
Copy link
Collaborator

@ptgoetz ptgoetz commented May 23, 2024

What?

Tighten up the RETRIEVAL_DESCRIPTION prompt to more reliably call retrieval with newer models like GPT 4 and 4o.

You have files uploaded as knowledge to pull from. Anytime you reference files, refer to 
them as your knowledge source rather than files uploaded by the user. You should adhere to the facts in the provided 
materials. Avoid speculations or information not contained in the documents. Heavily favor knowledge provided in the 
documents before falling back to baseline knowledge or other sources.

Why?

In testing retrieval with newer models like GPT 4 and GPT 4o, I noticed that those models can be reluctant to call retrieval in cases when they probably should.

Steps to reproduce:

  1. Create a new RAG or Assistant bot.
  2. Upload the OpenGPTs README.md (you may need to rename it to README.txt for the upload to be accepted. until my mime types PR is merged.)
  3. Save the Assistant.
  4. Ask "What is OpenGPTs?"

GPT 3.5 Turbo (it's the only one that actually calls retrieval):

OpenGPTs is an open source effort to create a similar experience to OpenAI's GPTs and Assistants API. It is powered by LangGraph, a framework for creating agent runtimes, and builds upon LangChain, LangServe, and LangSmith. OpenGPTs gives you more control over the language model, prompts, tools, vector database, retrieval algorithm, and chat history database used in your application. It also allows customization of the cognitive architecture, with support for assistants, RAG, and chatbots.

You can find more information and resources related to OpenGPTs in the provided links:

- [LangGraph](https://github.com/langchain-ai/langgraph)
- [LangChain](https://github.com/langchain-ai/langchain)
- [LangServe](https://github.com/langchain-ai/langserve)
- [LangSmith](https://smith.langchain.com/)

If you are interested in using OpenGPTs, you can follow the quickstart guide with Docker for setting up and running the project.

GPT 4 and later:

"OpenGPTs" does not refer to a specific, widely recognized technology or framework as of the last update. It's possible that you might be referring to "OpenAI's GPTs," which are a series of language models developed by OpenAI, including GPT (Generative Pre-trained Transformer), GPT-2, GPT-3, and the latest versions like GPT-4.

These models are designed to understand and generate human-like text based on the input they receive. They are used in a variety of applications, from chatbots and virtual assistants to more complex tasks like summarization, translation, and content generation.

If "OpenGPTs" refers to something specific in your context, could you please provide more details or clarify?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant