Examples
Worked examples
- Is an instance
LoRA fine-tuning of Llama 3 on a corpus of cardiology guidelines to build a domain assistant
Counter-examples
Looks similar, but isn't
- Not an instance
Few-shot prompting (showing examples in the prompt) is not fine-tuning — model weights are unchanged
Editorial commentary
Fine-tuning is the process of further training a pre-trained foundation model on a smaller, task-specific or domain-specific dataset, updating some or all of its parameters, to specialise its behaviour while retaining the general capability learned during pre-training. Common methods include full-parameter supervised fine-tuning (SFT), parameter-efficient methods such as LoRA (which train a small set of additional weights rather than the whole model), and preference-optimisation methods (RLHF, DPO) that adjust a model’s outputs toward human-preferred responses rather than toward a fixed labelled dataset.
Why the method matters, not just the fact of fine-tuning
“This model was fine-tuned” is an incomplete disclosure. The method determines what changed: full-parameter SFT can alter behaviour broadly, LoRA constrains the change to a narrow, cheaply reversible subspace, and RLHF/DPO change output preferences rather than factual knowledge. A fine-tune lineage record is the artefact that should capture which method was used, on what data, with what hyperparameters.
Why it matters for research
A fine-tuned model built on licensed or restricted training data inherits use restrictions from that data even where the base model itself is unrestricted — a distinction that matters directly for whether the fine-tuned model, or research results derived from it, can be shared or published.
References
- See also: RLHF, Large language model (LLM), Model lineage.
Also known as
Model adaptation · Domain adaptation (LLM)
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="Fine-tuning"
vocab-term-identifier="https://casrai.org/dictionary/term/fine-tuning" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/fine-tuning",
"name": "Fine-tuning",
"identifier": "https://casrai.org/dictionary/term/fine-tuning",
"description": "The process of further training a pre-trained foundation model on a smaller, task-specific or domain-specific dataset, updating some or all parameters, to specialise its behaviour while retaining general capability.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/genai-disclosure#set",
"url": "https://casrai.org/dictionary/term/fine-tuning",
"alternateName": [
"Model adaptation",
"Domain adaptation (LLM)"
],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"author": {
"@id": "https://casrai.org/#editorial-team"
},
"datePublished": "2026-05-21T01:57:28",
"dateModified": "2026-09-05T14:21:53",
"inLanguage": "en-GB",
"isAccessibleForFree": true
}
