AI engineering glossary · RAG vector search
What is similarity threshold?
Similarity threshold: A minimum similarity score a result must reach to be returned. Using it lets a RAG system return nothing, and say it does not know, when no stored passage is close enough, instead of passing weak matches to the model.
Similarity threshold: an example
Chunks with cosine similarity below 0.35 are dropped, so an off-topic question gets no context and the bot declines.
Related terms
More RAG vector search terms
- Small-to-big retrieval
- Sparse retrieval
- Structure-aware chunking
- Synthetic test set
- Table extraction
- Text extraction
- Top-k retrieval
- Two-stage retrieval
Where you learn it
Similarity threshold is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.