AWS Certified Cloud Practitioner

Explanation

Zero-shot prompting is the correct approach because:

1. Definition: Zero-shot prompting involves asking a model to perform a task without providing any examples or demonstrations. The model relies solely on its pre-trained knowledge and understanding of the task description.

2. Scenario Fit: The engineer wants the model to "answer without any prior examples and rely only on the model's general knowledge" - this is the exact definition of zero-shot learning.

3. Other options explained:

   • Code-generation prompting: This is for generating code, not classification tasks.

   • Retrieval-augmented prompting: This involves retrieving external information or examples to augment the prompt, which contradicts the requirement of "no prior examples."

   • Reinforcement learning prompting: This refers to training models through reward-based learning, not a prompting style for inference.

4. AWS Bedrock Context: In AWS Bedrock, zero-shot prompting is commonly used for classification tasks where you want the model to categorize inputs based on its general understanding without providing labeled examples.

Key Takeaway: When you need a model to perform a task based solely on its pre-existing knowledge without any examples, zero-shot prompting is the appropriate technique.