What is Context Stuffing?
AI EngineeringLast updated:
The practice of filling the LLM context window with maximum relevant information to improve response quality.
Context stuffing provides the model with extensive reference material, examples, and instructions. While it can improve accuracy, diminishing returns and the lost-in-the-middle problem mean strategic selection of context is more effective than simply maximizing volume.