Answer: Few-shot Prompting

## Explanation **Few-shot Prompting** is the correct answer because the developer is providing **two example inputs and outputs** before the main query. ### Key Concepts: 1. **Few-shot Prompting**: - Involves providing a small number of examples (typically 2-5) to demonstrate the desired task - Helps the model understand the pattern or format expected - The examples serve as a "demonstration" of how to process similar inputs 2. **Zero-shot Prompting**: - Would involve asking the model to perform the task without any examples - The model relies solely on its pre-trained knowledge 3. **Negative Prompting**: - Typically refers to specifying what NOT to include in the output - Common in image generation models 4. **Guided Prompting**: - Not a standard term in prompt engineering - May refer to providing step-by-step instructions ### Why B is Correct: The question explicitly states "They provide two example inputs and outputs before the main query," which is the definition of few-shot prompting. The examples help the model understand the classification task (positive vs. negative reviews) by showing it how similar inputs should be processed. ### Real-world Application: In AWS services like Amazon Bedrock or SageMaker, few-shot prompting is commonly used when working with foundation models to improve accuracy on specific tasks without full fine-tuning.

Author: Ritesh Yadav