Training & Learning
Fine-Tuning
Taking a pretrained model and training it further on a smaller, focused dataset to specialize it for a task or style.
Fine-tuning adapts an existing model to a narrower purpose by continuing its training on a targeted set of examples. It is far cheaper than training from scratch because the model already knows the basics and only needs to adjust. Teams fine-tune to teach a model a specific tone, domain knowledge, or format, and the instruction-following behavior of chatbots comes largely from a fine-tuning step.
Reference: Wikipedia: Fine-Tuning
Related terms
Pretraining
The first, largest training phase where a model learns general patterns from a huge amount of broad data.
Transfer Learning
Reusing what a model learned on one task as a head start for a different but related task.
Reinforcement Learning from Human Feedback
A training method that tunes a model to be helpful by rewarding the responses human reviewers prefer.
Foundation Model
A large, general-purpose model trained on broad data that many different applications can be built on top of.