Retrieval Quality/Latency

Kind: quality-attribute

Layer: Performance Core

Record: lexicon:retrieval-quality-latency

Canonical: Lexicon

The degree to which retrieval must trade result quality against speed.

Listed in Lexicon terms, after Monitoring Cost and before Prompt Registry.

Category

• Model Architecture

In tension with

• Retrieval-Augmented Generation
• Retrieval-Augmented Generation (RAG) / Retrieval Quality/Latency

Linked from

• Model Architecture
• The resolutions