Giving a model access to your information.
THE DEFINITION
Searching your documents for relevant passages, then giving the model those passages with the question.
Frequently abbreviated as RAG.
THE PROBLEM IT SOLVES
A model knows only what it learned during training. It does not know your documents.
WHY IT IS BETTER THAN FINE-TUNING FOR KNOWLEDGE
Information stays in your documents and can be updated without retraining The model can cite what it used Hallucination is substantially reduced
WHERE IT IS USED
Internal knowledge assistants Customer support answering from documentation Document question-answering
WHAT STILL GOES WRONG
If retrieval finds the wrong passages, the answer is wrong The model may add unsupported detail
WHY IT MATTERS TO YOU
Most practical business applications of language models use this approach.
RELATED TERMS
Embedding, fine-tuning, hallucination.
