Lesson 39 · 3 min
How fine-tuning works
Fine-tuning takes a model that already knows language and trains it a bit more on your own examples, so it picks up a style or a task.
Large models start as foundation models, trained on a huge amount of general text. They can do many things, but none of them exactly the way your team does.
Fine-tuning means training that model further on a small, focused set of examples: maybe a few hundred support replies in your company's tone, or forms filled in the right format.
After fine-tuning, the model leans toward that style without needing long instructions every time. It is good for consistent tone and format, and less good for teaching new facts, which change and are better looked up with RAG.
It needs clean examples, some technical work and testing afterwards. Many teams find that better prompts or a few examples in the prompt get them most of the way first.
Do it now
Plan a fine-tune on paper
- 1. Pick a repetitive writing task at work, like replying to a common customer question.
- 2. Ask a chatbot the prompt below to see whether fine-tuning would even help.
- 3. Collect five real examples of great answers. Even without fine-tuning, these make strong examples for a prompt.
I want an AI to handle this task: [DESCRIBE]. Would better prompts, examples in the prompt, document search or fine-tuning be the best fit? Explain the trade-offs simply.
Quick check
Keep going
Next →Open vs closed modelsWords from this lesson
- Fine-tuning
- Extra training of an existing model on your own examples so it gets better at one narrow kind of task.
- Training data
- The examples, such as text, images or recordings, that a model learns its patterns from.
- Foundation model
- A large general-purpose model trained on broad data that many different products are then built on top of.