Fine-tuning
Fine-tuning is the process of continuing the training of a pre-trained model on a smaller, task-specific dataset to specialise its behaviour.
What is Fine-tuning?
Fine-tuning updates some or all of a model's weights using supervised examples. Variants include full fine-tuning, parameter-efficient fine-tuning (LoRA, QLoRA), instruction tuning, and reinforcement learning from human feedback. Fine-tuning is most useful for fixed style, format, or domain vocabulary that prompting cannot reliably capture. It is rarely the right answer for adding new factual knowledge, where retrieval-augmented generation is cheaper, more auditable, and easier to update.
How does Fine-tuning apply to enterprise AI?
Enterprises fine-tune to enforce tone-of-voice, regulated output formats, multilingual code-switching, or domain jargon. Fine-tuning data must be governed under GDPR and the EU AI Act, including consent, retention, and the right to erasure.
Related terms
- Foundation Model - A foundation model is a large neural network pre-trained on broad data and designed to be adapted to many downstream tasks.
- RAG (Retrieval-Augmented Generation) - Retrieval-Augmented Generation (RAG) is an architecture pattern that grounds a language model's output in retrieved source documents rather than relying on the model's parametric memory alone.
- Prompt Engineering - Prompt engineering is the practice of designing, testing, and versioning the instructions given to a language model to elicit reliable, evaluable outputs.
- Data Residency - Data residency is the requirement that personal or regulated data stays within a specified geographic region throughout processing, storage, and backup.
External references
Need help applying Fine-tuning to your enterprise? Submit a short brief and we reply within one business day.