description: Prompt template you can build with examples.

Few Shot Prompt Template

The Few-Shot Prompt Template Function is a powerful tool for constructing prompts that guide AI models by providing them with a set of example input-output pairs - known as "few-shot examples." This technique is widely used to improve the accuracy and reliability of language models, especially for tasks requiring structured outputs or specific formats.

What Is Few-Shot Prompting?

• Few-shot prompting involves supplying the language model with a handful of demonstrations (examples) of the desired task within the prompt itself.

• These examples teach the model how to respond to new, similar queries by showing the expected pattern or structure

When to Use

• When you want to guide a language model with clear, structured demonstrations.

• For tasks where output format consistency is important.

• In scenarios where zero-shot prompting (no examples) yields unreliable results.

Parameters

Inputs

1. Examples (Required)

• Type: string (JSON)

• Description: Array of example objects with key-value pairs

• Example:

  [
    { "word": "happy", "antonym": "sad" },
    { "word": "tall", "antonym": "short" }
  ]```
  
  

2. Example Prompt (Required)

• Type: PromptTemplate

• Description: Template for formatting individual examples

3. Prefix (Optional)

• Type: string

• Description: Text appearing before the examples

• Example: “Give the antonym of every input”

4. Suffix (Required)

• Type: string

• Description: Text after examples, containing input variable

• Example: “Word:

\nAntonym:”

5. Example Separator (Optional)

• Type: string

• Default: “\n\n”

• Description: String used to separate examples

6. Template Format (Optional)

• Type: options

• Options: [“f-string”, “jinja-2”]

• Default: “f-string”

• Description: Format style for template strings

Best Practices

1. Example Design

   • Clear demonstrations

   • Consistent formatting

   • Diverse examples

   • Progressive complexity

2. Template Structure

   • Compatible formats

   • Clear separators

   • Proper variables

   • Consistent style

3. Performance Optimization

   • Efficient examples

   • Minimal redundancy

   • Clear instructions

   • Optimal spacing

{% hint style="info" %} This section is a work in progress. We appreciate any help you can provide in completing this section. Please check our Contribution Guide to get started. {% endhint %}