Question 19
Domain 2: Core Machine Learning, AI, and Transformer Foundations**Learning paradigm question:** In the context of LLMs, what does "few-shot learning" refer to?
Correct answer: B
Explanation
Few-shot learning means an LLM can adapt to a new task from only a small number of examples provided in the prompt. This matches the definition of learning with "minimal examples in the prompt," rather than requiring extensive retraining or large labeled datasets.
Why each option is right or wrong
A. Training models with very few parameters
B. Learning new tasks with minimal examples in the prompt
In the standard machine-learning taxonomy, “few-shot” denotes adaptation from only a small number of labeled demonstrations, typically just a handful of examples, rather than a large training set or parameter updates. In LLM prompting, this is implemented by placing a few task examples directly in the prompt so the model infers the pattern from context; the defining feature is the minimal number of examples, not retraining or fine-tuning.
C. Fast inference speed optimization
D. Processing very short text sequences