Fine-tuning is adapting a pre-trained model to a specific task or domain by continuing training on a smaller, targeted dataset — the most common way to specialize a general model. Full fine-tuning updates all parameters, while parameter-efficient methods like LoRA and adapters modify only a small subset. Fine-tuning can improve task performance dramatically but risks catastrophic forgetting of pre-trained knowledge if not managed carefully. The quality and diversity of the fine-tuning dataset often matters more than its size.