You give the model no examples, only an instruction. What is this?
Why B is the answer
Zero-shot prompting
Zero-shot uses no examples.
In detail
- Only an instruction, with no examples, is zero-shot.
- Add examples if results are off.
Use case
An illustrative scenario, not a real incident.
A writer asks for a product summary with no examples, which is zero-shot prompting.
See it in different ways
flowchart LR A["Zero-shot: no examples"] --> B["Single-shot: one example"] B --> C["Few-shot: several examples"] C --> D["Chain of thought: reason step by step"]
mindmap
root((Prompting))
Zero-shot
Instruction only
Few-shot
Examples
Chain of thought
Step by step
Negative prompt
What to avoidflowchart TD
Q{"How many examples?"}
Q -->|"None"| A0["Zero-shot"]
Q -->|"One"| A1["One-shot"]
Q -->|"Several"| A2["Few-shot"]Key point: Zero-shot uses no examples.
Why the other options are wrong
- A. Transfer learning
- Transfer learning reuses a model.
- C. Few-shot prompting
- Few-shot uses examples.
- D. Fine-tuning
- Fine-tuning changes a model.