Claude 편집하기 (부분)

==Defining Scope and Limitations in AI Instructions==
* Setting Clear Boundaries for Effective AI Interaction
When working with AI models, it’s crucial to clearly define the scope of the task and any applicable limitations. This practice helps ensure that the AI’s output aligns with your expectations and requirements while also preventing misunderstandings or overreach in the AI’s responses.


* Importance of Defining Scope and Limitations
 # '''Focus''': Helps the AI concentrate on the specific task at hand.
 # '''Accuracy''': Improves the relevance and precision of the AI’s output.
 # '''Efficiency''': Reduces time spent on irrelevant or out-of-scope information.
 # '''Expectations Management''': Aligns the AI’s capabilities with user expectations.
 # '''Ethical Considerations''': Prevents the AI from venturing into potentially problematic areas.

* Key Elements in Defining Scope and Limitations
 # '''Task Description''': Clearly state what the AI should do.
 # '''Context''': Provide relevant background information.
 # '''Boundaries''': Specify what should not be included or considered.
 # '''Data Sources''': Indicate what information the AI should use or ignore.
 # '''Output Format''': Define the desired format and structure of the response.
 # '''Time or Resource Constraints''': Mention any limitations on processing time or data usage.
 # '''Ethical Guidelines''': Outline any moral considerations or restrictions.

* Best Practices for Defining Scope and Limitations
 # '''Be Specific''': Use clear, unambiguous language to describe the task and its boundaries.
 # '''Prioritize''': Indicate the most important aspects of the task.
 # '''Use Examples''': Provide examples of what’s in-scope and out-of-scope when possible.
 # '''Consider Edge Cases''': Anticipate and address potential grey areas or exceptions.
 # '''Iterative Refinement''': Be prepared to adjust scope and limitations based on initial results.

* Examples of Defining Scope and Limitations
'''Example 1: Content Creation Task'''

<pre>
Task: Write a blog post about the benefits of exercise.

Scope:
- Focus on general physical and mental health benefits
- Include at least 3 types of exercise (e.g., cardio, strength training, flexibility)
- Target audience: adults aged 25-45 with sedentary jobs

Limitations:
- Do not discuss specific exercise routines or techniques
- Avoid mentioning branded fitness programs or equipment
- Do not make claims about treating or curing medical conditions
- Keep the post between 800-1000 words

Additional Guidelines:
- Use layman's terms, avoiding complex medical jargon
- Include 2-3 reliable sources for any statistical claims
- Conclude with a call-to-action encouraging readers to start exercising
</pre>

'''Example 2: Data Analysis Task'''
<pre>
Task: Analyze customer purchase data to identify trends and provide recommendations.

Scope:
- Use the provided CSV file containing 12 months of sales data
- Focus on identifying patterns in product categories, seasonal trends, and customer demographics
- Provide insights on top-selling products and underperforming areas

Limitations:
- Only use the data provided in the CSV file; do not incorporate external data sources
- Do not include any personally identifiable information about customers
- Limit the analysis to the past 12 months; do not extrapolate future trends

Output Requirements:
- Produce a summary report (max 2 pages) with key findings
- Include up to 3 data visualizations (charts or graphs)
- Provide a list of 3-5 actionable recommendations based on the analysis

Ethical Considerations:
- Ensure all insights and recommendations comply with data privacy regulations
- Avoid any analysis that could lead to discriminatory practices
</pre>

'''Example 3: Code Review Task'''
<pre>
Task: Review the provided Python script for a web scraping tool.

Scope:
- Check for code efficiency and adherence to PEP 8 style guidelines
- Identify any security vulnerabilities, particularly in handling user inputs
- Suggest improvements for error handling and logging

Limitations:
- Focus on the Python code only; do not review any associated HTML or JavaScript
- Limit suggestions to those that don't require changing the overall structure of the script
- Do not add new features; focus on improving existing functionality

Specific Areas to Address:
1. Input validation
2. Error handling and exception management
3. Use of appropriate libraries and functions
4. Code comments and documentation
5. Performance optimization for large datasets

Output Format:
- Provide a bullet-point list of issues found, categorized by severity (Critical, Major, Minor)
- For each issue, include:
  * The relevant line number(s)
  * A brief description of the problem
  * A suggested fix or improvement
- Conclude with a summary of the script's overall quality and main areas for improvement
</pre>

[https://salesforcefromscratch.co.uk/defining-scope-and-limitations-in-ai-instructions/ 범위나 제한 주기]