Glossary · AI technology
RAG (Retrieval-Augmented Generation)
RAG retrieves documents related to a question, adds them to the model input and generates an answer grounded in them.
- By
- OSKIT team (Afformation)
- Updated
Definition
RAG retrieves documents related to a question, adds them to the model input and generates an answer grounded in them.
RAG · Retrieval-Augmented Generation
How it works
The name comes from a 2020 paper by Lewis et al. It lets models answer with recent or private information they were not trained on.
AI search engines use the same structure. A page that is not retrieved cannot be cited in the generated answer.
What to check
- 01The page is in the search index
- 02Each page answers one question
Where OSKIT handles it
Post Engine
Post Engine writes from each client’s fact sheet and blocks numbers that are not in it at the scoring step.
See the featureSee also: Embedding · Grounding · Citation (in AI answers)
Sources
- [1]Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks (arXiv:2005.11401)https://arxiv.org/abs/2005.11401
Summarize in an AI chat
Only the page URL is sent.
A free check shows your search and AI answer visibility
Enter your website address and see search engine and AI bot access and index status right away.
3 minutes · Free · No contract