You show a model two examples of the format you want. Which technique is this?
Why C is the answer
Few-shot prompting
Few-shot prompting provides examples. Zero-shot gives none.
In detail
- Two examples is few-shot prompting.
- More examples use more tokens.
Use case
An illustrative scenario, not a real incident.
A writer shows two sample summaries in the prompt, and the model follows that format.
See it in different ways
flowchart LR A["Zero-shot"] --> B["One-shot"] B --> C["Few-shot"] C --> D["Prompt chaining"] D --> E["Chain of thought and ReAct"]
mindmap
root((Prompting))
Examples
Zero-shot
One-shot
Few-shot
Persona
Role prompting
Steps
Prompt chaining
Tools
ReActflowchart TD
Q{"How many does the prompt give?"}
Q -->|"None"| A0["Zero-shot"]
Q -->|"One"| A1["One-shot"]
Q -->|"A few"| A2["Few-shot"]Key point: Few-shot prompting gives examples.
Why the other options are wrong
- A. Zero-shot prompting
- Zero-shot uses no examples.
- B. Role prompting
- Role prompting sets a persona.
- D. Prompt chaining
- Chaining links prompts.