RAG is only as good as the text in the store it retrieves from, and what’s retrieved is only useful as context.
KEG (Knowledge-Engineered Generation), on the other hand, enforces a much higher quality of reference context, its concepts can be used for governance, and it allows for more “human in the loop”.
See the article at Beyond RAG: Knowledge-Engineered Generation for LLMs
Great article but I have so many questions, like how do you get extra HITL using a KG?
For chat (say) The simplest way is to introduce a browsing function into the chat UI. The user can then explore the concepts and then choose one to ask a question about.
You can try it out by clicking up the beth chat icon to the right and typing “knowledge graph”. Browse by entering a number.
