Review Fine-Tuning

Fine-tuning is additional training applied to an existing model using selected examples or feedback to change its behavior for a task, domain, or response pattern.

Instead of building a model from the beginning, a team starts with a pretrained model and adjusts some of its parameters using a narrower training process.

Fine-tuning can improve consistent behavior, but it creates data, evaluation, maintenance, cost, and version-management obligations that prompting or retrieval may avoid.

A company fine-tunes a model on reviewed examples so it follows a specialized output format more consistently.

Fine-tuning is not the same as giving a model current documents to consult. Retrieval supplies information at use time; fine-tuning changes learned behavior.

Fine-tuning changes the model; retrieval changes what the model can consult now.

* [[foundation-model|Foundation Model]] * [[training-data|Training Data]] * [[retrieval-augmented-generation|Retrieval-Augmented Generation]] * [[weights|Weights]]

* [NIST AI 100-2 E2025, Adversarial Machine Learning](https://doi.org/10.6028/NIST.AI.100-2e2025) — NIST terminology for model training and fine-tuning. * [Google for Developers, Machine Learning Glossary](https://developers.google.com/machine-learning/glossary) — technical definition and examples of fine-tuning.

* Compare with [[retrieval-augmented-generation|Retrieval-Augmented Generation]] before choosing an adaptation method.