Claude 편집하기 (부분)

==Specifying Desired Output Format in AI Instructions==

* Tailoring AI Responses for Maximum Utility

When working with AI models, clearly specifying the desired output format is crucial for obtaining results that meet your needs. By providing explicit instructions about how you want the information presented, you can ensure that the AI’s response is accurate, immediately useful, and easy to integrate into your workflow.

* Common Output Formats
# '''Paragraphs''': For narrative explanations or detailed descriptions.
# '''Bullet Points''': For concise lists of items or key points.
# '''Numbered Lists''': For step-by-step instructions or prioritized information.
# '''Tables''': For organized presentation of data or comparisons.
# '''JSON''': For structured data that needs to be parsed by software.
# '''CSV''': This is for tabular data that can be easily imported into spreadsheets.
# '''Markdown''': This is for formatted text that can be easily converted to other formats.

* Best Practices for Specifying Output Format
 - '''Be Explicit''': Clearly state the format you want, e.g., “Present the information in a table with three  -  Date, Event, and Outcome.”
 - '''Provide Examples''': If possible, give a sample of the desired format to guide the AI.
 - '''Specify Details''': Include information about headers, column names, or specific formatting requirements.
 - '''Consider the End Use''': Consider how the output will be used and specify a format that aligns with that purpose.
 - '''Be Consistent''': Use similar format specifications across related tasks for uniformity.

* Examples of Format Specifications

'''Example 1: Data Analysis Report'''

<pre>
Generate a sales analysis report with the following sections:

1. Executive Summary (2-3 paragraphs)
2. Monthly Sales Breakdown (table format with columns: Month, Revenue, % Change)
3. Top 5 Products (bullet point list with product name and total sales)
4. Regional Performance (bar chart description in markdown format)
5. Recommendations (numbered list with 3-5 actionable items)

Ensure each section has a clear heading and subheadings where appropriate.
</pre>

'''Example 2: Technical Documentation'''
<pre>
Create API documentation in the following format:
# API Endpoint Name

## Description
[1-2 sentences describing the endpoint's purpose]

## Request Format
- Method: [GET/POST/PUT/DELETE]
- URL: `[endpoint URL]`

## Parameters
| Name | Type | Required | Description |
|------|------|----------|-------------|
| [param1] | [string/int/etc] | [Yes/No] | [Brief description] |
| [param2] | [string/int/etc] | [Yes/No] | [Brief description] |

## Response Formatjson
{
  "key1": "value1",
  "key2": "value2"
}## Error Codes
- 400: Bad Request
- 401: Unauthorized
- 404: Not Found
- 500: Internal Server Error

## Example Usage
[Provide a code snippet demonstrating how to use the API]

Repeat this structure for each API endpoint in the documentation.

</pre>

[https://salesforcefromscratch.co.uk/specifying-desired-output-format-in-ai-instructions/ 결과물 형태 제시]