What is fine-tuning?
Fine-tuning is further training of an existing model on a specific dataset, adjusting its weights so it behaves more like the examples it was shown.
Fine-tuning is often proposed as the way to make a model 'know' a business. It is usually the wrong tool for that, and the reason is about update cost. Facts change — a price rises, a policy is rewritten, a product is discontinued — and a fine-tuned model has to be retrained to learn the new version, with no way to remove the old one cleanly.
Why it matters
The useful split is this: fine-tuning teaches behaviour, retrieval supplies facts. If you need a model to answer in a particular structure every time, classify into your own categories, or adopt a house voice that prompting cannot reach, fine-tuning is genuinely the tool. If you need it to know today's shipping cost, retrieval is, because updating the answer then means editing a page rather than running a training job.
Examples
- Fine-tuning a model to always return a structured triage category — behaviour, and appropriate.
- Fine-tuning a model on last year's product catalogue — facts, and it will be wrong within a month.
- Retrieving today's pricing page and letting an untouched model read it — facts, correctly handled.
Not the same as
RAG
RAG changes what a model is shown at question time. Fine-tuning changes the model itself and cannot be audited per answer.
Prompting
Prompting is free, immediate and reversible. Try it exhaustively before concluding that fine-tuning is required.
Related
term
What is RAG (retrieval-augmented generation)?
RAG is what separates a chatbot that knows your business from one that guesses. Documents are split into passages and indexed by meaning; each question retrieves the closest passages, and the model is instructed to answer only from them.
term
What is a large language model?
The training objective is narrow — predict the next piece of text — and the capabilities that emerge from it are broad. A model that has read enough text learns grammar, tone, structure, and a great deal of general world knowledge, without ever being taught any of those things directly.
term
What is a system prompt?
The system prompt is where an operator says what the assistant is, what it may talk about, what to do when it does not know, and how to sound. It is the cheapest and fastest lever available — a change takes effect on the next request with no retraining and no deployment of anything else.
RubyRep is this, applied to your website.
Give it your URL and it reads your pages, then answers your customers from them.
Train My RubyRep