Definition
The process of dynamically inserting retrieved external data or state information into an LLM's prompt to ground its response in specific, real-time facts. While it significantly reduces hallucinations, architectural trade-offs include increased inference latency and higher token costs as the prompt size grows.
Distinguish from 'Prompt Injection' (a security vulnerability); Context Injection is a deliberate design pattern for data grounding.
"A lawyer being handed a specific case file right before a trial to ensure their arguments are based on current evidence rather than general legal knowledge."
- Retrieval Augmented Generation (RAG)(Parent Framework)
- Context Window(Physical Constraint)
- Grounding(Functional Objective)
- Tokenization(Prerequisite Process)
Conceptual Overview
The process of dynamically inserting retrieved external data or state information into an LLM's prompt to ground its response in specific, real-time facts. While it significantly reduces hallucinations, architectural trade-offs include increased inference latency and higher token costs as the prompt size grows.
Disambiguation
Distinguish from 'Prompt Injection' (a security vulnerability); Context Injection is a deliberate design pattern for data grounding.
Visual Analog
A lawyer being handed a specific case file right before a trial to ensure their arguments are based on current evidence rather than general legal knowledge.