LoRA (low-rank adaptation)

Also written: LoRA

A fine-tuning method that trains small added matrices while the model's own weights stay frozen. It is covered in a later book.