Evaluating a Model Adaptation Strategy for a Startup

A research team is adapting a large language model for a specialized task but has a very limited budget for computational resources. They decide to use a method where only a small set of newly introduced, task-specific vectors are trained, while the millions of original model parameters remain unchanged. Which statement best analyzes why this approach is computationally efficient?

A key reason that fine-tuning a model by only training a small set of new vectors prepended to each layer is computationally efficient is that this method inherently requires a much smaller training dataset compared to methods that update the entire model.