How Adaptable Are Finetuned Models?

How well are fine-tuned models able to adapt to new functions being added to their function calling stack? Would they need to be retrained?

1 Like