RAG that actually works?
The user is looking for a Retrieval Augmented Generation (RAG) system that can effectively utilize a large library of complex texts to provide accurate and comprehensive answers.
Why it matters
Developing effective RAG systems that can understand and utilize large, complex knowledge bases is crucial for creating truly capable AI assistants and knowledge management tools.
Key Points
- 1The user tried using AnythingLLM and llama.cpp to create a knowledge base, but found the results to be poor.
- 2When asking about a specific topic mentioned in a long PDF, the system claimed it could not find any mention of that topic.
- 3The user is not a developer, just a regular user, and is looking for an easy-to-use RAG system that can truly understand large, complex text libraries.
Details
The user is a regular person, not a developer, who tried to create a knowledge base using the AnythingLLM tool and llama.cpp. The goal was to have an 'expert' system that could provide comprehensive information on specific topics, similar to a human expert. However, the user found that the results were poor, with the system claiming it could not find mentions of topics that were clearly present in the included PDF documents. This suggests that the current RAG systems are not able to fully utilize and understand the information contained in large, complex text libraries. The user is looking for a RAG system that can truly comprehend and leverage such extensive knowledge sources to provide accurate and detailed answers.
No comments yet
Be the first to comment