Definition

Fine-Tuning

Fine-tuning is the targeted further training of an already pretrained AI model on a smaller, task- or domain-specific dataset to adapt its behaviour for a particular purpose.

Fine-tuning permanently bakes style, format or expertise into the model but requires training data and effort. For many Swiss SMEs, RAG or good prompting is initially cheaper; fine-tuning pays off for stable, recurring specialised tasks.

← Back to glossary