Definition
Fine-tuning starts from a pre-trained model and trains it further on a curated set of examples: inputs paired with the outputs you want. Variants include supervised fine-tuning, preference tuning and parameter-efficient methods such as LoRA, which change only a small part of the model.
Fine-tuning is good at teaching behaviour (a house style, a strict output format, a classification task) and at making a smaller model match a larger one on a narrow job. It is a poor way to add facts that change, which is what retrieval is for. It also creates a model you must version, evaluate and retrain when the base model is updated.