Definition
Prompt Augmentation is the process of dynamically enriching a base query by injecting external context, such as retrieved document snippets or system-level metadata, into the model's input. It bridges the gap between a generic user request and a grounded response, though it requires balancing increased token latency and cost against the quality of the generated output.
Distinguish from static prompt engineering; augmentation is a dynamic, data-driven enrichment step occurring at runtime.
"A legal brief being stuffed with relevant case-law exhibits right before a trial begins."
- Vector Database(Prerequisite)
- Context Window(Architectural Constraint)
- RAG (Retrieval-Augmented Generation)(Parent Framework)
- Few-shot Prompting(Component)
Conceptual Overview
Prompt Augmentation is the process of dynamically enriching a base query by injecting external context, such as retrieved document snippets or system-level metadata, into the model's input. It bridges the gap between a generic user request and a grounded response, though it requires balancing increased token latency and cost against the quality of the generated output.
Disambiguation
Distinguish from static prompt engineering; augmentation is a dynamic, data-driven enrichment step occurring at runtime.
Visual Analog
A legal brief being stuffed with relevant case-law exhibits right before a trial begins.