What is fine-tuning in AI?
Fine-tuning is the process of further training a pre-trained AI model on a smaller, task-specific set of examples to adapt its behavior, style or skills. It is useful for consistent formats and specialized tasks, while RAG is usually better for keeping facts current.
Fine-tuning vs RAG
| Fine-tuning | RAG | |
|---|---|---|
| Changes | Model behavior | Information given |
| Fresh data | Needs retraining | Immediate |
| Citations | No | Yes |
| Best for | Style, format, skills | Facts, policies |
When to fine-tune
Consistent output formats, domain language and classification tasks where prompts alone fall short.
