Fine-Tuning
Continuing to train a pre-trained model on domain-specific data to adapt its behavior.
Fine-tuning updates a foundation model's weights using a smaller, task-specific dataset. This specializes the model for a domain (medical, legal, coding) or enforces a particular output format. Fine-tuned models can outperform prompt engineering alone for narrow tasks, but they are expensive to create, lock in a model version, and require retraining when the base model updates.
Related Terms
The initial large-scale training phase where a model learns language from massive text corpora.
Fine-tuning a model on examples of instructions paired with ideal responses.
Reinforcement Learning from Human Feedback — training models to align with human preferences.
A large pre-trained model that serves as the base for many downstream applications.