Fine-tuning is the process of taking a general, pre-trained model and giving it extra training on a smaller, focused dataset so it performs better at a particular task or in a particular style. Instead of building a model from scratch — which is enormously expensive — you start from a capable base model and specialize it.
For example, a general language model might be fine-tuned on a company's support conversations so it answers in that company's tone and knows its products. Fine-tuning is one of several ways to adapt AI; for many everyday needs, a well-written prompt achieves the same goal without any retraining.
Related terms
Large Language Model (LLM) · Training Data · Prompt Engineering