Skip to content

Glossary: Fine-tuning LLMs

Adapter

Also written: adapters, LoRA adapter

The small set of trained LoRA matrices, saved as a separate file. It loads on top of the unchanged base model, can be swapped for another task, or can be merged into the weights.

Where the book teaches it

Related terms

LoRA (low-rank adaptation)Merging (an adapter)PEFT (parameter-efficient fine-tuning)