ModelRefs / Instruction Following — AI Glossary
Instruction Following — AI Glossary
A model's ability to accurately execute explicit user or system instructions, including format, length, style, and constraint compliance.
Overview
Instruction following is the primary capability unlocked by SFT and RLHF. Evaluated by IFEval (Google) which tests measurable constraints: word count limits, JSON format, inclusion/exclusion of strings. Frontier models score 85–95% on IFEval; small models and uncensored models often score <70%.
Reference details
| Topic | prompting |
|---|---|
| Last reviewed | 2026-06-24 |
Related terms
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Instruction Following — AI Glossary.
Frequently asked questions
What is Instruction Following?
A model's ability to accurately execute explicit user or system instructions, including format, length, style, and constraint compliance.
What concepts are related to Instruction Following?
Closely related concepts include sft, system message, calibration.