Few-Shot Learning
Definition
A model's ability to learn a new task from just a few examples provided in the prompt.
Few-shot learning provides a small number of input-output examples in the prompt to demonstrate the desired behavior. The model uses these examples to infer the pattern and apply it to new inputs. This is more effective than zero-shot learning for tasks where the desired output format or style is not easily described in words.
In text refinement, few-shot learning might involve showing the model two or three examples of raw transcriptions paired with their desired refined versions. The model then applies the same transformation pattern to new transcriptions. Ummless presets can leverage few-shot examples for more precise control over refinement behavior.