Category:
Category:
Category:
RAG & Retrieval
Definition
Automatically inserting retrieved or generated context into LLM prompts.
Explanation
Context injection ensures the LLM receives the right supporting information before generating an output. It is used in RAG systems, agent workflows, and knowledge assistants. Context can come from retrieval, memory, tool outputs, or agent collaboration. Effective context injection dramatically reduces hallucinations and increases task accuracy.
Technical Architecture
Query → Retrieval/Memory → Context Packager → LLM → Output
Core Component
Retriever, context window allocator, packager, prompt formatter
Use Cases
Knowledge assistants, legal QA, enterprise copilots
Pitfalls
Irrelevant or excessive context reduces accuracy; token wastage
LLM Keywords
Context Injection, Prompt Context, RAG Context
Related Concepts
Related Frameworks
• RAG
• Chunking
• Context Window
• Context Assembly Pipeline