RAG & Search
Dense retrieval
Retrieval that uses learned vector representations to find semantically similar content.
Example
A meaning-based search represents both questions and documents with dense vectors.
Why it matters
It helps connect related ideas that exact word matching may miss.
You'll hear it
“The search found a useful passage without the exact phrase.”
Can you control it?
Set by the team
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Can it overlook rare names or codes?
- Yes. Specialized identifiers may need exact matching alongside meaning-based comparisons.
- Does dense mean the source documents are lengthy?
- No. It describes the numerical form used for searching, not the writing style or document length.
- Can changes in the learned descriptions affect results?
- Yes. A different preparation method can rearrange which items appear similar.