You add two example inputs and outputs to a prompt. What is this called?
Why C is the answer
Few-shot prompting
Microsoft explains that few-shot examples guide the current response only. The model is not permanently changed.
In detail
- Examples in a prompt are few-shot prompting.
- It needs no training at all.
Use case
An illustrative scenario, not a real incident.
A team adds two sample inputs and outputs to a prompt, and the model copies the format. That is few-shot prompting.
See it in different ways
sequenceDiagram participant App as Your app participant M as Model App->>M: System prompt with role and rules App->>M: User prompt M-->>App: Response
mindmap
root((Settings))
Temperature
Lower is more consistent
Higher is more varied
Max tokens
Limits reply length
System prompt
Role and rulesflowchart LR S0["Example 1"] --> S1["Example 2"] --> S2["New input"] --> S3["Matching output"]
Key point: Examples in a prompt are few-shot.
Why the other options are wrong
- A. Fine-tuning a new model
- Fine-tuning changes a model, and this does not.
- B. Zero-shot prompting
- Zero-shot uses no examples.
- D. Model training
- Nothing is being trained.