Context Extension

Definitions

1. An approach for increasing the amount of context a model can handle, whether through architectural changes, fine-tuning methods, retrieval, or system-level techniques. Context extension matters when tasks require longer documents, histories, or memory than the base model supports comfortably.

   In plain English: A way of letting an AI system work with longer context.

   Example: "The team pursued context extension so legal review prompts could include full contract sections instead of fragmented excerpts."

Related Terms

• Context Length
• Data Mixture
• Dense Model
• Frontier Model
• Internal Representation
• Large Context Model
• Likelihood
• Model Alignment
• Model Fingerprint
• Model Gateway
• Needle in a Haystack
• Preference Model
• Sparse Model
• Student Model