F
AI CONCEPTS
Fine-tuning
Also known as: Transfer Learning, Domain Adaptation
Definition
Fine-tuning adapts a pre-trained LLM to a specific task or domain by training it further on specialized data. This produces a model that's more accurate and aligned with your specific use case without training from scratch.
Simple Explanation
A pre-trained LLM is like a well-educated generalist. Fine-tuning is like sending them to graduate school for your specific field. They keep their general knowledge but become experts in your domain.
Real-World Use Case
A legal firm fine-tunes an LLM on their internal case law and contract templates. The resulting model writes in-house contracts much more accurately than a general LLM.
Examples
LoRAQLoRAInstruction tuningDomain-adaptive pre-training
Related Terms
Learn More
- OpenAI Fine-tuning Guide
- Hugging Face Training