Fine-tuning

Fine-tuning is the process of taking a pre-trained large language model (LLM) and further training it on a smaller, specific dataset to specialize its performance.

Unlike prompt engineering, which happens at runtime, fine-tuning modifies the model's internal weights. This is useful for:

• Style Adaptation: Forcing the model to speak in a specific brand voice or format.
• Domain Expertise: Teaching the model industry-specific terminology (e.g., medical or legal).