Fine-tuning
The process of continuing the training of a pre-trained model on specific data to specialize it for a task or domain.
Fine-tuning takes a pre-trained model (like a base LLM) and continues training it on a smaller, task-specific dataset. The result is a model that's specialized for that task while retaining the general knowledge from pretraining.
Common fine-tuning use cases include adapting a general LLM to medical, legal, or customer-support domains; teaching a model your company's tone of voice; or training a model to follow specific output formats.
Modern fine-tuning techniques include LoRA (Low-Rank Adaptation), which trains only a small number of new parameters rather than the entire model — making fine-tuning faster and cheaper. For most use cases in 2026, prompt engineering and RAG achieve similar results without fine-tuning, but fine-tuning still wins when consistent style or format matters.