RAG
A technique for grounding that finds relevant text in your documents or data and adds it to the model's context, so the model writes its answer from that text without being changed.
Also called
- Retrieval-augmented generation
Kinds
- Vector database A type of database that stores embeddings and can find the closest match to a query by meaning.
The other half
- Fine-tuning A further round of training that adapts a pretrained model to specific tasks or a particular domain.
Related
- Knowledge cutoff The latest date used in the data that trains a model.
- GEO (generative engine optimization) A marketing practice to shape content so that AI search and answer tools are more likely to find it and cite it.
How it connects
- RAG adds sources to Context
- Grounding usually done with RAG
- Knowledge base searched by RAG
- Embedding power the search in RAG
- RAG is the method behind Retrieval from our data