← Back to the A-Z glossary
AI glossary · P2

Fine-tuning

Fine-tuning adapts an existing model on examples to improve one specific behavior.

Customization 4 min read Updated 2026-05-22
— Definition

Fine-tuning, in plain words

Fine-tuning adapts an existing model on examples to improve one specific behavior.

Put fine-tuning back in its place: powerful, but not your first move.

A concrete example

You feed it thousands of already-validated classification examples so the model reproduces that labeling more reliably.

Why it matters

It can sharpen a style or a repeatable task, but it costs time and needs good data.

You see it on AI platforms, open-weight models, and projects with lots of high-quality examples.

Don't mix it up with

RAG: RAG lets an AI answer using documents pulled up at the moment you ask the question.

Evaluation / Eval: An eval regularly measures whether an AI answers correctly on representative cases.

Common mistakes

  • Fine-tuning to add knowledge that changes often.
  • Using a handful of weak examples.
  • Forgetting to measure before/after with an eval.

Quick checklist

  • First I check whether the word names a concept, a tool, a risk, or a metric.
  • I tie it to a concrete case: You feed it thousands of already-validated classification examples so the model reproduces that labeling more reliably.
  • I keep the main trap in mind: Fine-tuning to add knowledge that changes often.

Quick questions

What is Fine-tuning in AI?

Fine-tuning adapts an existing model on examples to improve one specific behavior.

Where will I run into Fine-tuning?

You see it on AI platforms, open-weight models, and projects with lots of high-quality examples.

Which word should I read next?

Start with RAG, Evaluation / Eval, LoRA / QLoRA.

Want to keep going in order?

Head back to the full glossary, search a word, then open only the pages that deserve more than a short definition.

Open the AI glossary