Fine-tuning
Further training a pre-trained model on specific data to adapt it.
Fine-tuning changes the model’s weights to improve performance on a particular task or style. It needs controlled, high-quality data and re-testing, because it can also introduce bias or remove safety behaviours.
On the exam
A fine-tuned model is a new model version: it needs change management, testing and documentation before deployment.
Related terms
- Foundation modelA large general-purpose model that can be adapted to many tasks.
- Change management for AIControlling changes to models, data, prompts, thresholds and settings.
- Retrieval-augmented generation (RAG)Giving a language model relevant documents at the time of the question, so answers are grounded in them.
AAIA Test yourself on this with timed AAIA practice questions in the AAIA Exam Simulation Hub, or read the free AAIA exam guide.
Test yourself with today’s question of the day.
