WebsitePlatform Login

Expert-Level Prompting

Master-level prompting techniques for highest precision and performance

These advanced techniques are intended for experienced users who want to get the maximum out of AI systems.

Constitutional AI Prompting

Self-Critique and Improvement

Task: [Your request]

After formulating an initial response:
1. Evaluate your answer according to these criteria:
   - Factual accuracy
   - Completeness
   - Clarity and comprehensibility
   - Practical applicability
   
2. Identify areas for improvement

3. Create a revised, optimized version

4. Present only the final, improved answer

Ethics Framework Integration

Before answering, check:
- Fairness: Is the answer balanced and unbiased?
- Harm: Could the information be misused?
- Transparency: Are uncertainties clearly marked?
- Benefit: Does the answer maximize value for the user?

Adjust your response accordingly.

Recursive Prompting

Self-Expanding Prompts

Step 1: Analyze this request and identify 3 sub-questions
that need to be answered.

Step 2: For each sub-question:
- Formulate it precisely
- Answer it thoroughly
- Identify other relevant aspects

Step 3: Synthesize all partial answers into a coherent
overall response.

Step 4: If important new questions have emerged, repeat
the process once.

Depth-First Exploration

Topic: [Your topic]

1. Choose the most important aspect
2. Explore it to complete depth
3. Document all insights
4. Return and choose the next aspect
5. Repeat until all aspects are covered
6. Create an integrated summary

Adversarial Prompting

Counter-Argumentation

Thesis: [Your position]

1. Formulate the strongest argument FOR this thesis
2. Switch perspective: Formulate the strongest COUNTER-arguments
3. Analyze both sides objectively
4. Identify synthesis opportunities
5. Present a nuanced conclusion

Stress Testing

Proposed solution: [Your idea]

Test this solution against:
1. Edge cases
2. Scaling problems
3. Unforeseen side effects
4. Resource constraints
5. Human factors

For each identified problem: Suggest adjustments

Prompt Programming

Functional Prompts

DEFINE FUNCTION analyze_text(text, parameters):
    aspects = extract_aspects(parameters)
    results = {}
    
    FOR EACH aspect IN aspects:
        results[aspect] = deep_analyze(text, aspect)
    
    RETURN synthesize(results)

END FUNCTION

# Execution
analyze_text(
    text="[Your text]",
    parameters={
        "aspects": ["sentiment", "key_points", "implications"],
        "depth": "comprehensive",
        "output_format": "structured"
    }
)

Conditional Logic

IF input_type == "technical":
    use_technical_language = True
    detail_level = "high"
ELSE IF input_type == "general":
    use_technical_language = False
    detail_level = "medium"
    add_examples = True
ELSE:
    ask_for_clarification()

Generate_response_with_parameters()

Emergent Behavior Prompting

Creativity Maximization

Think outside all conventional boundaries:

1. Initially ignore all practical constraints
2. Combine concepts from completely different domains
3. Question fundamental assumptions
4. Develop 5 radically innovative approaches
5. Only then: Adapt the most promising for reality

Synergy Discovery

Elements: [A, B, C]

Examine:
1. A + B: What synergies emerge?
2. B + C: What new possibilities?
3. A + C: What unexpected connections?
4. A + B + C: What emergent properties?
5. Identify the synergy maximum

Multi-Modal Reasoning

Cross-Domain Transfer

Problem in Domain A: [Description]

1. Abstract the core problem
2. Find analogous problems in Domains B, C, D
3. How were they solved there?
4. Transfer solution principles back to Domain A
5. Adapt for specific context

Metaphorical Thinking

Complex concept: [Your concept]

1. Find metaphors from:
   - Nature (biological systems)
   - Technology (machines, software)
   - Society (social structures)
   - Art (music, painting)

2. For each metaphor:
   - What are the parallels?
   - What insights does it offer?
   - Where are the limits?

3. Synthesize a new, powerful explanation

Advanced Output Control

Probabilistic Thinking

Question: [Your question]

Answer structure:
1. Most likely answer (70-90% confidence)
2. Alternative scenarios (20-30% confidence)
3. Unlikely but possible cases (5-10%)
4. Black Swan events (<1% but high impact)

For each scenario: Justification and implications

Granular Control

Parameters for this response:
- Technical level: [1-10]
- Example density: [low/medium/high]
- Cultural context: [specify]
- Time horizon: [short/medium/long-term]
- Risk tolerance: [conservative/balanced/progressive]

Generate tailored response based on these parameters.

Meta-Learning Prompts

Prompt Optimization

Original prompt: [Your prompt]

1. Analyze strengths and weaknesses
2. Generate 3 improved versions:
   - Version A: Clearer structure
   - Version B: More precise requirements
   - Version C: Better examples
3. Test each version
4. Combine the best elements
5. Present optimal prompt

Adaptive Prompting

After each response evaluate:
- Was the answer too detailed/superficial?
- Was the desired style achieved?
- Were there misunderstandings?

Adjust the next prompt accordingly:
- For too little detail: "Explain in more detail..."
- For wrong style: "In the style of..."
- For misunderstandings: "To clarify..."

Advanced Applications

Complex Decision Making

Decision problem: [Description]

Framework:
1. Stakeholder analysis (Who is affected?)
2. Criteria definition (What is important?)
3. Option generation (What is possible?)
4. Impact analysis (What are the consequences?)
5. Risk assessment (What can go wrong?)
6. Synergy check (How to combine?)
7. Temporal consideration (Short vs. long term)
8. Decision matrix
9. Sensitivity analysis
10. Final recommendation with confidence level

Innovation Framework

Innovation goal: [Your goal]

Phase 1 - Divergence:
- Brainstorming without limits
- Cross-industry inspiration
- Future scenarios
- Wild combinations

Phase 2 - Convergence:
- Feasibility assessment
- Resource analysis
- Market potential
- Implementation path

Phase 3 - Synthesis:
- Top 3 innovation concepts
- Detailed elaboration
- Prototyping plan
- Success metrics

Achieving Mastery

True mastery in prompting comes through:

  1. Continuous experimentation - Try new techniques
  2. Systematic learning - Document successes and failures
  3. Creative combination - Mix techniques situationally
  4. Deep understanding - Understand AI functionality
  5. Ethical reflection - Responsible use

Further Resources

On this page