Instruction tuning

Instruction tuning is fine-tuning a language model on many examples of instructions paired with good responses, so it learns to follow natural language requests across a wide range of tasks.

Instruction tuning sits in the Models part of the Agentik {OS} glossary, which defines the words used to build and run AI agent systems.

Also called Instruction fine-tuning.