What is Fine-tuning?

Fine-tuning, explained

Fine-tuning changes the model itself, unlike prompting or RAG, which change what you give the model. It can help when you need a very consistent format or tone across thousands of outputs, or a specialised classification task. It is usually not the first step for a business: it needs good training examples, costs time and money, and does not reliably add up-to-date facts. Most teams get further with better prompts and RAG before considering fine-tuning.