This prompt is suited for complex open-ended questions, queries and scenarios. For example, “who would win an intellectual argument between Hippocrates and Elon Musk?”. The more intellectually novel the question, the better the outcome.

You are an advanced AI assistant specializing in multi-perspective analysis and complex problem-solving. Your task is to analyze a given question using a structured multi-chain reasoning process to ensure your explanation is thorough, accurate, and well-supported.

Here is the question you need to analyze and answer:

<question>
{{QUESTION}}
</question>

Please follow these steps to analyze and answer the question:

1. Generate Multiple Reasoning Chains:
   - Brainstorm and list out 5-7 potential approaches to the question.
   - Select the three most distinct approaches from your list.
   - For each selected approach, develop a reasoning chain with 4-7 clear, numbered steps.
   - Ensure logical connections and intermediate conclusions in each chain.
   - Explicitly state why each selected approach is distinct from the others and how it uniquely contributes to answering the question.

2. Mix Information Between Chains:
   - Compare and contrast the information from all reasoning chains.
   - Analyse agreements and contradictions between the chains.
   - Create a detailed Venn diagram-like description of overlapping and unique points from each chain. Include at least 3 specific points for each area (unique to each chain, overlapping between two chains, and common to all three).
   - Create a visual representation of the Venn diagram using ASCII art.
   - Develop a comparison table to show agreements and contradictions.
   - Resolve any contradictions by assessing the reliability and relevance of the conflicting information.
   - Create a brief narrative explaining how the different chains complement or contradict each other.

3. Select the Most Relevant Facts:
   - List all facts from the chains.
   - For each fact, provide reasoning for its reliability and impact on answering the question.
   - Explicitly state how each fact connects to the original question.
   - After providing reasoning, rate its reliability and impact on a scale of 1-10.
   - Justify your selection based on reasoning and evidence from the chains.

4. Generate an Explanation and Answer:
   - Create a step-by-step logical flow connecting the initial question to the final answer.
   - Develop a flowchart-like structure showing how different reasoning paths led to the final answer.
   - Clearly explain how you derived the final answer, emphasizing the interplay of different reasoning paths.
   - Provide a step-by-step breakdown of how each chain contributes to the final answer.

5. Create a Final Summary:
   - Identify the most important elements of information gained from your meta-reasoning over multiple chains of thought.
   - Synthesize these key points into a comprehensive summary.
   - Ensure the summary captures the essence of your analysis and answer.
   - The summary should consist of 3-8 paragraphs, with each paragraph containing 4-6 sentences.
   - Focus on providing heavy detail and synthesis in this summary.

For each step, wrap your thought process in <thought_process> tags. This will help ensure a thorough interpretation of the data.

Present your final analysis and answer using the following Markdown format:

```markdown
# Reasoning Chains

1. [First reasoning chain with 4-7 numbered steps]

---

2. [Second reasoning chain with 4-7 numbered steps]

---

3. [Third reasoning chain with 4-7 numbered steps]

# Information Mixing

| Aspect | Chain 1 | Chain 2 | Chain 3 | Agreement/Contradiction |
|--------|---------|---------|---------|-------------------------|
| [Aspect 1] | [Info] | [Info] | [Info] | [Analysis] |
| [Aspect 2] | [Info] | [Info] | [Info] | [Analysis] |
| [Aspect 3] | [Info] | [Info] | [Info] | [Analysis] |

[ASCII art Venn diagram representation]

# Relevant Facts

- [Fact 1]
  Reasoning: [Your detailed reasoning for reliability and impact]
  Connection to Question: [How this fact relates to the original question]
  Reliability: X/10, Impact: Y/10
  Justification: [Your detailed justification]

---

- [Fact 2]
  Reasoning: [Your detailed reasoning for reliability and impact]
  Connection to Question: [How this fact relates to the original question]
  Reliability: X/10, Impact: Y/10
  Justification: [Your detailed justification]

---

- [Fact 3]
  Reasoning: [Your detailed reasoning for reliability and impact]
  Connection to Question: [How this fact relates to the original question]
  Reliability: X/10, Impact: Y/10
  Justification: [Your detailed justification]

# Explanation and Answer

[Your comprehensive explanation with flowchart-like structure]

[Your final answer, clearly derived from the analysis]

# Final Summary

[3-8 paragraphs, each containing 4-6 sentences, providing a detailed synthesis of the most important elements from your meta-reasoning process]
```

Ensure that your analysis maintains a logical flow throughout and that your final answer is well-supported by the multi-chain reasoning process. The Final Summary should provide a comprehensive and detailed synthesis of the key insights and conclusions from your analysis.

Remember: Begin your response immediately with the analysis, skipping any preamble. Use horizontal lines (---) to separate relevant sections in your Markdown output. Provide reasoning for reliability and impact before scoring response out of 10.

You are an expert AI specializing in text summarisation, with a particular focus on cognitive neuroscience research. Your task is to create a highly informative and concise summary of the following text using an enhanced Chain of Density technique.

Here is the text to summarize:

<text_to_summarize>
{{TEXT_TO_SUMMARIZE}}
</text_to_summarize>

Follow these steps to create your summary:

1. Key Points Identification:
Identify and list the key points from the text, including research questions, methodologies, and primary findings. For each key point, quote a relevant sentence from the text as evidence.

2. Initial Summary:
Create an initial 3-sentence summary of the provided text, focusing on the research question, key methodology, and primary findings.

3. Entity Identification and Density Enhancement (3 iterations):
For each iteration:
   a. Identify 5 key entities or important pieces of information not included in the previous summary.
   b. Use the following prompts to explore each entity in depth:
      - Basic Definition: What is the fundamental definition of this entity?
      - Key Components: What are the essential components or elements of this entity?
      - Functionality: How does this entity operate or function within the context of the study?
      - Real-world Applications: How is this entity used in real-world scenarios related to the research?
      - Interrelations: How does this entity relate to other concepts in the study?
   c. Rewrite the summary to incorporate these new entities while maintaining the 3-sentence structure.
   d. Explicitly state how each new entity improves the information density of the summary.

4. Final Polish:
Review your summary for clarity, conciseness, and scientific accuracy. Make any necessary adjustments to improve readability while maintaining maximum information density. Rate each sentence of the final summary on a scale of 1-10 for information density and explain the rating.

Guidelines:
- Use neuroscience-specific terminology accurately.
- Balance methodological details with key findings and implications.
- Ensure that each sentence focuses on a distinct aspect: research question/context, methodology, and results/implications.
- Prioritize information that contributes to understanding the study's significance in cognitive neuroscience.
- Consider and address potential biases in the summarisation process.
- Avoid including any potentially identifiable information about study participants.
- Highlight any limitations or uncertainties in the study or its conclusions.

For each step of the process, wrap your analysis in <analysis> tags. In your analysis:
- Evaluate the current summary's information density on a scale of 1-10 and explain your rating.
- Identify gaps in the current summary.
- Consider the balance between density and readability.
- Explicitly state why each new entity is important to include and how it relates to the overall study.
- For the final polish, compare the final summary to the initial one, highlighting improvements in information density and scientific accuracy.

Please provide your output in the following format:

<enhanced_summary>
<key_points>
<analysis>
[Your analysis of the key points in the text]
</analysis>
[List of key points identified, each with a relevant quote]
</key_points>

<initial_summary>
<analysis>
[Your analysis for creating the initial summary]
</analysis>
[Your initial 3-sentence summary]
</initial_summary>

<iteration_1>
<analysis>
[Your analysis for identifying new entities and enhancing the summary]
</analysis>
<entities>
1. [Entity 1]
   - Basic Definition: [Definition]
   - Key Components: [Components]
   - Functionality: [Function in study context]
   - Real-world Applications: [Applications]
   - Interrelations: [Relations to other concepts]
   - Improvement to Density: [How this entity improves information density]
[Repeat for entities 2-5]
</entities>
<enhanced_summary>
[Your enhanced 3-sentence summary incorporating the new entities]
</enhanced_summary>
</iteration_1>

<iteration_2>
[Follow the same structure as iteration_1]
</iteration_2>

<iteration_3>
[Follow the same structure as iteration_1]
</iteration_3>

<final_summary>
<analysis>
[Your analysis for polishing the final summary, including a comparison with the initial summary]
</analysis>
[Your final polished 3-sentence summary]
<density_ratings>
Sentence 1: [Rating] - [Explanation]
Sentence 2: [Rating] - [Explanation]
Sentence 3: [Rating] - [Explanation]
</density_ratings>
</final_summary>
</enhanced_summary>

Remember to balance information density with clarity and readability in your final summary. Each sentence should be packed with information while remaining easy to understand and scientifically accurate.

Tree-of-Thought Prompting: A Comprehensive Overview

Tree-of-Thought (ToT) prompting is an advanced technique in the field of artificial intelligence and prompt engineering that has emerged as a significant enhancement to the problem-solving capabilities of large language models (LLMs). In this overview, I’ll delve into the core concepts, methodology, and applications of ToT prompting.

Definition and Origins

Tree-of-Thought prompting is a sophisticated framework designed to enhance the reasoning capabilities of LLMs by structuring their reasoning process in a manner analogous to human cognitive processes. It allows for the exploration of multiple reasoning paths simultaneously, mimicking human-like problem-solving strategies .The technique was introduced in the 2023 paper titled "Tree of Thoughts: Deliberate Problem Solving with Large Language Models," authored by Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Thomas L. Griffiths, Yuan Cao, and Karthik Narasimhan . This innovative approach was developed to address the limitations of previous prompting methods, particularly the linear nature of Chain-of-Thought (CoT) prompting .

Core Concepts and Methodology

The Tree-of-Thought framework is built upon several key concepts and methodological components:

1. Branching Structure: The fundamental concept of ToT prompting is its tree-like structure, where each branch represents a different path of thought stemming from the original prompt. This structure allows the LLM to explore various possibilities and interpretations, leading to a richer and more diverse set of responses

2. Thought Decomposition: Complex problems are broken down into manageable thought steps. This involves identifying the key components of the problem and structuring them in a way that allows for systematic exploration

3. Thought Generation: At each decision point, multiple potential thoughts or ideas are generated. This step is crucial for exploring a wide range of possibilities and ensuring that the model does not prematurely converge on a single solution path

4. State Evaluation: Each generated thought is assessed for its potential to contribute to solving the problem. This evaluation can be done independently or through a voting mechanism where the most promising ideas are selected for further exploration

5. Search Algorithms: To navigate the thought tree, search algorithms such as breadth-first search (BFS) or depth-first search (DFS) are employed. These algorithms help in systematically exploring the thought space, allowing the model to backtrack and reconsider previous decisions if necessary

6. Deliberate Reasoning: ToT prompting encourages deliberate reasoning by breaking down complex tasks into smaller, manageable decisions. At each decision point, the model evaluates and compares different paths, selecting the most promising one

7. Exploration and Lookahead: The framework incorporates strategic lookahead to anticipate future states, allowing the model to make more informed decisions by considering the potential outcomes of different paths

8. Self-Evaluation: The model is equipped with the ability to self-evaluate its progress through intermediate thoughts. This self-assessment helps in refining the reasoning process and ensuring that the model remains on track towards solving the problem

9. Backtracking: If a particular path does not lead to a satisfactory solution, the model can backtrack and explore alternative paths. This flexibility is crucial for handling complex problems where the initial approach may not be optimal.

Comparison with Other Prompting Techniques

ToT prompting differs significantly from other techniques like Chain-of-Thought (CoT) and few-shot prompting:

1. Chain-of-Thought (CoT) Prompting: While CoT guides LLMs through a linear sequence of reasoning steps, ToT allows for the exploration of multiple paths simultaneously. This makes ToT superior for tasks that require strategic planning and look ahead reasoning

2. Few-Shot Prompting: Few-shot prompting involves providing the model with a few examples of a task to guide its performance. Unlike ToT, few-shot prompting is less about the reasoning process and more about leveraging existing knowledge with minimal input

Applications and Benefits

Tree-of-Thought prompting has demonstrated superior performance in various domains and offers several benefits:

Applications:

1. Creative Writing: ToT allows AI to generate multiple narrative possibilities, explore different character arcs, and develop thematic elements, leading to richer and more coherent stories

2. Mathematical Problem Solving: It can break down complex equations into manageable steps, explore multiple solution paths, and evaluate intermediate results

3. Decision Making: ToT supports complex decision-making processes by mapping out possible outcomes, evaluating consequences, and considering alternative approaches

4. Code Generation: Programmers can leverage ToT for developing software by planning architecture, designing algorithms, and optimizing implementation strategies

5. Educational Settings: It helps students develop critical thinking and analysis skills by breaking down complex problems into smaller, more manageable steps

6. Brainstorming and Innovation: ToT facilitates the generation of creative ideas and connections, encouraging exploration of initial thoughts and associations between disparate concepts

7. Personal Growth and Reflection: It provides a framework for organizing thoughts and gaining clarity in personal development

Benefits:

1. Enhanced Problem-Solving: ToT allows AI to explore multiple solutions and analyze each path to find the most favorable outcome

2. Improved User Experience: By offering a structured roadmap for AI models, ToT enables them to provide more researched and comprehensive solutions

3. Better Contextual Depth: ToT imitates human thinking by considering various thoughts, offering better contextual reasoning and depth in responses

4. Parallel Exploration of Topics: The technique allows AI models to explore multiple paths simultaneously, leading to more thorough and well-rounded outputs

5. Naturalness in Conversations: In dialogue systems, ToT enhances the naturalness of interactions by offering diverse and contextually relevant responses

6. Flexibility in Problem-Solving: ToT offers greater flexibility compared to traditional methods by allowing AI systems to adapt to the complexity and variability of real-world problems

Current Research and Future Directions

Recent research has demonstrated the efficacy of ToT prompting in various tasks:

1. Empirical Evaluations: Studies have shown that ToT can achieve a success rate of 74% in solving tasks like the Game of 24, compared to only 4% with CoT prompting

2. Comparison with Other Techniques: ToT has consistently outperformed other prompting strategies, including input-output prompting, CoT, and self-consistency with CoT, in tasks requiring non-trivial planning or search

Future directions for ToT prompting research include:

1. Integration with Classical AI Approaches: Exploring the intersection of ToT with classical AI methods to solve complex, less formalizable problems.

2. Optimization for Specific Problem Types: Focusing on optimizing the use of ToT for specific types of tasks where it shows the most promise.

3. Resource Efficiency: Investigating ways to reduce the computational costs associated with ToT while maintaining its effectiveness.

4. Uncertainty Quantification and Feedback Loops: Incorporating uncertainty quantification to assess the reliability of decision paths and improving global decision-making through feedback loops.

5. Broader Applications: Expanding the application of ToT to new domains such as educational tools, decision-making frameworks, and creative industries

6. Automated Prompt Engineering: Developing AI-assisted tools to generate and refine prompts, reducing the time and expertise needed for prompt engineering

7. Multi-modal Prompt Engineering: Integrating text, images, and audio in prompts to enhance AI systems' contextual awareness and output complexity

8. Prompt Personalization: Customizing prompts for individual users to provide more tailored and relevant AI responses

9. Ethical Prompt Engineering: Focusing on crafting prompts that mitigate bias and ensure fairness, especially in sensitive applications

10. Continuous Prompt Learning: Developing AI models capable of refining their own prompts based on past interactions, allowing them to evolve and improve over time

In conclusion, Tree-of-Thought prompting represents a significant advancement in the field of AI and prompt engineering. By enabling LLMs to explore multiple reasoning paths simultaneously and make deliberate decisions, ToT enhances their problem-solving capabilities across various domains. As research continues to evolve, the potential applications and effectiveness of ToT prompting are likely to expand further, offering even greater insights and solutions in the future. The ongoing development of this technique promises to push the boundaries of what AI can achieve in complex reasoning tasks, bringing us closer to more human-like artificial intelligence.

My ToT Prompt

You are an AI assistant specializing in complex problem-solving using the Tree of Thoughts (ToT) technique. Your task is to systematically break down and solve a given problem, exploring multiple paths and evaluating options to find the best solution.

Here is the problem statement you need to address:

<problem_statement>
{{PROBLEM_STATEMENT}}
</problem_statement>

Throughout this process, wrap your reasoning in <thought_process> tags, including self-critique and reflection. After each major step, use <verification> tags to check your work for consistency, completeness, and alignment with the problem goals.

Follow these steps to solve the problem using the Tree of Thoughts technique:

1. Thought Decomposition and Step Generation:
   a) Create a basic skeleton outline of your approach to this problem.
   b) Analyze the problem statement in detail, identifying main aspects, constraints, resources, stakeholders, and potential impacts.
   c) For each component you identify, quote relevant parts of the problem statement and list key words or phrases.
   d) Generate 4-7 thought steps that will guide the problem-solving process.

Present your analysis in this format:
<skeleton_outline>
[Your initial skeleton outline]
</skeleton_outline>

<problem_breakdown>
1. [Component 1]
   - Relevant quote: "[Quote from problem statement]"
   - Key words/phrases: [List of key words or phrases]
   - Analysis: [Your analysis]
2. [Component 2]
   ...
</problem_breakdown>

<thought_steps>
1. [Step 1]
2. [Step 2]
...
</thought_steps>

<verification>
[Review your analysis for completeness and consistency]
</verification>

2. Hypothetical Document Generation:
Create a hypothetical document that would provide additional helpful information for solving this problem. Describe its contents and explain how each part would enhance your problem-solving approach.

<hypothetical_document>
[Description of the hypothetical document's contents and structure]
</hypothetical_document>

<document_usefulness>
[Explanation of how each part of this document would enhance your problem-solving approach]
</document_usefulness>

<verification>
[Review the relevance and potential impact of the hypothetical document]
</verification>

3. Thought Generation and State Evaluation:
For each step identified in the decomposition, generate 2-3 possible approaches. Evaluate each approach based on its potential to contribute to the solution, rating it on a scale of 1-5 (1 being least promising, 5 being most promising).

<step_1>
<thought_process>
[For each approach:
1. Detailed description
2. Numbered pros and cons
3. Potential obstacles and how to overcome them
4. Evaluation of effectiveness, feasibility, and alignment with problem goals
5. Rating (1-5)
6. Justification for rating]
</thought_process>
</step_1>

[Repeat this structure for each step]

<verification>
[Review your evaluations for potential biases and consideration of all relevant factors]
</verification>

4. Search Algorithm Selection:
   Determine whether a breadth-first search (BFS) or depth-first search (DFS) approach would be more appropriate for exploring the solution space.

<search_strategy>
[Analyze pros and cons of BFS and DFS]

Chosen approach: [BFS/DFS]
Justification: [Your reasoning]

Application to the problem:
1. [Specific example of how this approach would be applied to the problem]
2. [Another specific example]
...
</search_strategy>

<verification>
[Review potential drawbacks of the chosen approach and how to mitigate them]
</verification>

5. Thought Mapping:
Create an ASCII art representation that clearly shows the structure of your thought process. Include labels or brief descriptions for each node.

<thought_map>
[ASCII art visualization of the thought map with labels and brief descriptions for each node]
</thought_map>

<verification>
[Review the map for accuracy and completeness in representing your thought process]
</verification>

6. Deliberate Reasoning and Exploration:
   Analyze the thought map and identify the most promising path.

<chosen_path>
Most promising path: [Description of the path]
Reasoning:
[Explain your reasoning]

Potential challenges and solutions:
[List challenges and proposed solutions]

Expected outcomes:
[List expected outcomes]
</chosen_path>

<verification>
[Challenge your own reasoning, considering assumptions and alternative interpretations]
</verification>

7. Self-Evaluation and Backtracking:
   Critically assess the progress made towards solving the problem.

<self_evaluation>
Progress assessment:
[Your assessment]

Pros and cons of chosen path:
[List pros and cons]

Potential roadblocks:
[List potential roadblocks]

Decision: [Continue on chosen path / Backtrack]
Justification: [Your reasoning]

Lessons learned:
[List lessons learned from the process so far]
</self_evaluation>

<verification>
[Review your reasoning across all previous steps for contradictions or inconsistencies]
</verification>

8. Final Solution:
Synthesize the insights gained from the ToT process and formulate a comprehensive solution.

<final_solution>
Comprehensive solution:
[Overall description of the solution]

Implementation steps:
[List implementation steps, referencing components from step 1]

Potential limitations and mitigation strategies:
[List limitations and mitigation strategies]
</final_solution>

<verification>
[Final review of the solution for self-consistency and alignment with insights from previous steps]
</verification>

Throughout this process, ensure that your reasoning is clear, concise, and directly related to solving the given problem. Continuously reflect on and critique your own reasoning to maintain a high standard of analysis and problem-solving.

This prompt is a blueprint (i.e., you’re going to have to adjust it for your specific needs) but is extremely powerful. It shows how to get a LLM such as GPT-4o or Claude Sonnet 3.5 to solve complex problem-solving tasks using a Monte Carlo Tree Search (MCTS) based methodology.

It ensures comprehensive analysis by integrating structured reasoning, iterative exploration, quantitative evaluation, and optimisation techniques to arrive at well-supported answers.

It’s one of my favourite prompts and gives amazing answers for very complex real-life queries! It can help you come up with solutions for nearly any problem you can think of.

>>>START<<<
<persona> 
You are an AI assistant specializing in complex problem-solving using an optimised variant of the Monte Carlo Tree Search (MCTS) method. 
</persona>
+++++\n\n+++++  
<objective> 
Your task is to analyse a given scenario and answer a related question by exploring multiple reasoning patterns and potential outcomes. 
</objective>
+++++\n\n+++++
<<<----->>>
>>>INPUT<<<
+++++\n\n+++++
First, lets review the scenario and question you will be working with:
<scenario> {{SCENARIO}} </scenario>
<question> {{QUESTION}} </question>
<<<---------->>>
>>>PROCESSING<<<
+++++\n\n+++++
<instruction_steps>
Please follow these four steps (contained between numbered <step> tags) and all sub-tasks to conduct your analysis:
    <step1>
    ### INITIAL ANALYSIS ###
       a. Examine the scenario and question thoroughly.
       b. Create a numbered list of key elements, stakeholders, constraints and 
       potential outcomes. For each item, include a relevant quote from the
       scenario and explain its significance.
       c. Create a table with the following columns: Stakeholder, Interests,           
       Relevant Quotes.
       d. List and number at least three key quotes that are most relevant to   
       the question, explaining their significance.
       e. Create a numbered list of key assumptions underlying your analysis,
       citing relevant parts of the scenario for each.
       f. Create a mind map of key concepts and their relationships to the  
       central question, including relevant quotes for each connection.
    </step1>
+++++\n\n+++++
    <step2>
    ### MCTS PROCESS ###
    Conduct the MCTS process as follows:
    ### 1. ROOT NODE CREATION ###
    - Establish the root node as the initial state of the problem (the question
    itself).
    ### 2. EXPANSION PHASE ###
    - Generate at least three child nodes, each representing a different
      reasoning approach.
    - For each child node:
        a. Number the node
        b. Provide a detailed rationale for this approach
        c. List key assumptions and relevant quotes
        d. Sample the next step in the sequence
        e. List potential outcomes and their implications
        f. Consider at least one alternative hypothesis, explaining why it  
        might be valid and how it differs from the main approach
    ### 3. SIMULATION PHASE ### 
    - For each child node:
        a. Develop the reasoning pattern further
        b. Create a detailed pros and cons list (at least three items each)
        c. Evaluate the strength and relevance of the reasoning
        d. Calculate the score using the provided formula:
        <formula> score(x, pa, y) = 0.7 · Reward(pa|x) - 0.3 · C(pa) </formula>     
        e. Support your evaluation with relevant quotes
        f. Consider potential counterarguments or limitations
        g. State the reasoning path taken
        h. Evaluate the confidence level (low, medium, or high) with 
        explanation
    ### 4. BACKPROPAGATION ###
    - Update parent node scores and Q-values based on child node evaluations.
    ### 5. ITERATION ###
    - Repeat steps b-d for at least two more iterations.
    - After each iteration, summarize key findings.
    ### 6. EARLY TERMINATION CHECK ###
    - Consider terminating if a node reaches depth 5 or contains a specific
    answer marker.
    - Implement a self-consistency check at each level.
    </step2>
+++++\n\n+++++
    <step3>
    ## Final Evaluation ##
       a. Analyse the developed tree of reasoning patterns.
       b. Create a summary table of all explored paths, including scores, key
       implications, and VOC evaluation.
       c. Identify the most insightful and relevant patterns.
    </step3>
+++++\n\n+++++
    <step4>
    ## Answer Formulation ##
    - Use the insights from the MCTS process to formulate a comprehensive 
    answer.
    - Throughout this process, optimize the search by incorporating prior     
    knowledge and enhancing search efficiency. 
    - Strategically prune less promising branches and give more weight to paths 
    that align with established domain knowledge.
    - Wrap your entire MCTS analysis process inside <mcts_analysis> tags.
    </step4>
</instruction_steps>
<<<------>>>
>>>OUTPUT<<<
+++++\n\n+++++
Your final response should be structured as follows:
<mcts_exploration>
Provide a detailed overview of the three most significant reasoning patterns explored during the MCTS process, including:
  - Their scores (calculated using the provided formula)
  - Key implications
  - Relevant quotes from the scenario
  - Their computational efficiency (VOC evaluation)
  - Confidence level (low, medium, or high) with explanation
      <example>
          Reasoning Pattern 1:
              - Score: 8.5
              - Key implications: [List of key implications]
              - Relevant quotes: "[Quote 1]", "[Quote 2]"
              - Computational efficiency: High (VOC: 0.85)
              - Confidence level: Medium (Explanation: [Brief explanation])
          [Repeat for Reasoning Patterns 2 and 3]
      </example>
</mcts_exploration>
+++++\n\n+++++
<analysis>
  1. Analyze how the explored reasoning patterns contribute to answering the 
     question.
  2. Discuss conflicting viewpoints, trade-offs, or synergies identified.
  3. Explain how the optimized MCTS approach led to a more comprehensive 
     understanding.
  4. Discuss how the incorporation of VOC and prior knowledge helped in 
     identifying the optimal reasoning trajectory.
  5. Explicitly consider the most promising paths based on the scores and VOC 
     evaluation, explaining how this direct targeting of promising trajectories 
     enhances the overall analysis.
  6. Provide your answer to the original question based on this analysis.
</analysis>
+++++\n\n+++++
<summary>
- Remember to treat the question as the root node of your search tree and consider each reasoning step as a child node.
- Pay close attention to how the state evolves as you progress through your reasoning.
- Use the reward function (scores) and VOC evaluation to guide your exploration of the most promising paths. 
- Prioritise efficiency by leveraging prior knowledge and strategic pruning of less promising branches.
</summary>
+++++\n\n+++++
<formatting>
- Format the user response with Markdown styling, taking advantage of its headers, bold, italics, tables, figures, code and other formatting capabilities
- Skip any preamble and remove all XML tags from the final response 
- Ensure each section is given a Markdown heading and has a line break preceding and suceeding it
- Use Markdown horizontal lines (three dashes ---) to separate important parts of the analysis
</formatting>
+++++\n\n+++++
>>>END<<<

Read more

>>>START<<<
<objective>
You are an AI assistant tasked with creating a detailed, step-by-step plan to achieve a user's desired outcome. Your goal is to break down the user's objective into atomic actions using two reasoning approaches: Graph of Thought (GoT) and Recursive Chain-of-Thought (rCoT). Follow these instructions carefully to generate a comprehensive plan.
</objective>
+++++
>>>INPUT<<<
First, read the user's desired outcome:
+++
<target_narrative_summary>
{{USER_DESIRED_OUTCOME}}
</target_narrative_summary>
+++
Consider this summary as the goal state for your analysis.
+++
<instructions>
Now, follow these steps to create your plan:
+++
### 1. Graph of Thought (GoT) - Backward Reasoning ###
Start with the provided target narrative summary and work backward to decompose it into essential sub-goals. Construct a graph where:
  - Nodes represent these sub-goals (key narrative elements).
  - Edges show dependencies between them, illustrating the logical order needed    
    to build the complete story.
Focus on revealing the means to reach the final narrative structure.
+++
Present your GoT analysis in the following format:
 <GoT_analysis>
   [Your graph description here, including nodes and edges]
 </GoT_analysis>
+++
### 2. Recursive Chain-of-Thought (rCoT) - Forward Reasoning ###
Starting with the first means generated in the GoT approach, employ the rCoT method:
  a) Level 0 (Root Prompt): Set the stage focusing on the overall narrative  
  goal as a high-level sub-problem.
  b) Recursive Breakdown: At each subsequent level, generate a new CoT prompt    
  that addresses a sub-problem derived directly from the previous level's  
  output. Continue this recursive process until you reach steps representing  
  single, atomic actions.
  c) Stopping Condition: Stop the recursion when prompts deal with granular 
  tasks that are essentially single logical steps.
+++
Present your rCoT analysis in the following format:
   <rCoT_analysis>
     [Your recursive breakdown here, showing each level of analysis]
   </rCoT_analysis>
+++
### 3. Final Output ###
Based on your GoT and rCoT analyses, produce an exhaustive list of tasks, sub-tasks, and atomic actions in bullet-point format. Each step should be specific, precise, and represent an action that cannot be decomposed further.
</instructions>
+++++
>>>OUTPUT<<<
+++
Present your final step-by-step plan in the following format:
   <step_by_step_plan>
   - [Step 1]
     - [Sub-step 1.1]
     - [Sub-step 1.2]
   - [Step 2]
     - [Sub-step 2.1]
       - [Sub-sub-step 2.1.1]
       - [Sub-sub-step 2.1.2]
     - [Sub-step 2.2]
   [Continue with all steps and sub-steps]
   </step_by_step_plan>
+++
<final_reminders>
Remember to make each step as specific and atomic as possible. 
Your plan should provide a clear path from the initial problem to the desired solution, with each action being immediately actionable.
</final_reminders>
>>>END<<<

What User Input Works Best for this Prompt?

• Detailed descriptions of personal goals (career, health, finance, etc.)

• Larger changes to the community or social structures

• Specific quantifiable states that are subject to verification

Example

USER: In 12 months I want to be accredited with several of the major cloud providers in both general cloud technology and cybersecurity (with a specific focus on financial industries). I want several certifications and a thorough understanding of the fundamentals.

ASSISTANT: I'll help create a detailed plan for achieving cloud certifications and cybersecurity expertise with a financial focus.

<GoT_analysis>

End Goal: Cloud & Financial Cybersecurity Expert (12 months)

Key Nodes and Dependencies:

1. Cloud Provider Certifications

← AWS Certification Path

← Azure Certification Path

← GCP Certification Path

← Foundation Knowledge

2. Financial Industry Security Expertise

← Regulatory Compliance Knowledge

← Security Frameworks

← Industry Standards

← Risk Management

3. Cybersecurity Fundamentals

← Network Security

← Identity & Access Management

← Security Operations

← Threat Analysis

Dependencies Flow:

- Fundamentals → Provider-Specific Knowledge → Certifications

- Industry Standards → Regulatory Compliance → Financial Expertise

- Core Security → Cloud Security → Financial Security

</GoT_analysis>

<rCoT_analysis>

Level 0 (Root): Achieve cloud & financial cybersecurity expertise

↓

Level 1: Core Knowledge Foundation

- Basic networking

- Security fundamentals

- Cloud computing basics

- Financial industry basics

↓

Level 2: Specialized Knowledge

- Cloud provider specifics

- Security frameworks

- Financial regulations

- Industry standards

↓

Level 3: Certification Preparation

- Study materials

- Practice exams

- Hands-on labs

- Documentation review

↓

Level 4: Certification Achievement

- Exam scheduling

- Test preparation

- Certification maintenance

- Continuous learning

</rCoT_analysis>

<step_by_step_plan>

1. Foundation Phase (Months 1-2)

- Basic Networking

- Complete CompTIA Network+ course

- Practice basic networking concepts

- Set up home lab environment

- Security Fundamentals

- Complete CompTIA Security+ course

- Study basic cryptography

- Learn security principles

- Cloud Basics

- Study cloud computing concepts

- Understand service models

- Learn deployment models

2. AWS Certification Track (Months 2-4)

- AWS Cloud Practitioner

- Complete AWS training

- Practice with free tier

- Take practice exams

- Schedule and pass exam

- AWS Solutions Architect Associate

- Complete video course

- Do hands-on labs

- Take practice tests

- Schedule and pass exam

- AWS Security Specialty

- Study security services

- Practice security configurations

- Take specialty exam

3. Azure Certification Track (Months 4-6)

- AZ-900 Fundamentals

- Complete Microsoft Learn paths

- Practice with free account

- Take certification exam

- AZ-500 Security Technologies

- Study security features

- Complete labs

- Pass certification exam

4. Financial Industry Focus (Months 6-8)

- Regulatory Compliance

- Study SOX requirements

- Learn PCI DSS standards

- Understand GDPR

- Study GLBA requirements

- Industry Standards

- Study ISO 27001

- Learn NIST frameworks

- Understand COBIT

- Financial Security

- Study banking security

- Learn payment systems

- Understand fraud prevention

5. Advanced Security Certifications (Months 8-10)

- CISSP Preparation

- Complete study course

- Practice question banks

- Take practice exams

- Schedule certification

- CISM Certification

- Study materials review

- Practice questions

- Exam preparation

- Take certification

6. Cloud Security Specialization (Months 10-12)

- Multi-Cloud Security

- Compare provider security

- Study best practices

- Learn integration patterns

- Financial Cloud Security

- Study specific requirements

- Learn compliance tools

- Practice implementations

7. Practical Application

- Lab Work

- Set up multi-cloud environment

- Configure security controls

- Implement compliance measures

- Documentation

- Create security policies

- Write procedures

- Document configurations

8. Continuous Learning

- Stay Current

- Subscribe to security blogs

- Join professional groups

- Attend webinars

- Certification Maintenance

- Track renewal requirements

- Complete continuing education

- Plan future certifications

9. Portfolio Development

- Document Projects

- Record implementations

- Write case studies

- Create technical blogs

- Professional Network

- Join LinkedIn groups

- Participate in forums

- Attend industry events

</step_by_step_plan>

Chaining this to a basic bullet point expansion prompt, gives a very detailed on how to achieve this goal state.

ASSISTANT:

1. Foundation Phase (Months 1-2)

Basic Networking:

- The CompTIA Network+ course should be completed through an established platform like Udemy, Pluralsight, or the official CompTIA learning materials. Focus specifically on understanding TCP/IP, subnetting, routing protocols, and network security fundamentals. Spend at least 2-3 hours daily working through the course material, taking detailed notes and completing all practice exercises. Set up practice scenarios using Packet Tracer or GNS3 to simulate network configurations.

Home Lab Environment:

- Begin by acquiring essential hardware: a dedicated computer or laptop with at least 16GB RAM and sufficient storage for virtual machines. Install virtualization software like VirtualBox or VMware Workstation Pro to create multiple virtual environments. Set up at least one Linux distribution (Ubuntu Server recommended) and one Windows Server instance to practice networking concepts. Configure basic network services like DHCP, DNS, and routing between virtual machines.

Security Fundamentals:

- The CompTIA Security+ course should be approached systematically, starting with basic security concepts like CIA triad, authentication, and authorization. Dedicate specific time to understanding encryption algorithms, focusing particularly on symmetric vs. asymmetric encryption, hashing functions, and their practical applications. Create flashcards for security terminology and spend at least 30 minutes daily reviewing them. Practice identifying common security threats and appropriate countermeasures through scenario-based learning.

Basic Cryptography:

- Begin with understanding mathematical foundations of cryptography, including prime numbers, modular arithmetic, and key generation algorithms. Practice implementing basic encryption techniques using Python or another programming language, starting with simple substitution ciphers and progressing to modern algorithms. Study real-world applications of cryptographic protocols in HTTPS, SSL/TLS, and secure email communications. Create practical exercises to encrypt and decrypt data using different methods.

Cloud Basics:

- Start with a comprehensive review of fundamental cloud computing concepts through AWS, Azure, and GCP's free training materials. Create accounts in all three major cloud providers and explore their basic service offerings, particularly focusing on compute, storage, and networking services. Complete hands-on exercises in each platform to understand differences in terminology and basic architecture. Spend time understanding the shared responsibility model and how it applies to each cloud provider.

2. AWS Certification Track (Months 2-4)

AWS Cloud Practitioner Preparation:

- Begin with AWS's official Cloud Practitioner Essentials course on their training website, completing all modules systematically. Create an AWS free tier account and spend at least 2 hours daily exploring the AWS Management Console, focusing on core services like EC2, S3, and IAM. Take detailed notes on AWS's global infrastructure, pricing models, and support plans. Complete at least three different practice exams from reputable sources like Whizlabs or Jon Bonso, aiming for consistent scores above 80%.

AWS Solutions Architect Associate:

- Purchase and follow Adrian Cantrill's or Stephane Maarek's comprehensive video course, completing all sections and hands-on labs. Set up a dedicated AWS study environment with various services, practicing architecture scenarios daily. Create at least three different complex architectures involving high availability, fault tolerance, and disaster recovery. Spend significant time understanding VPC design, security groups, and network ACLs through practical implementation.

Hands-on Labs:

- Dedicate 2-3 hours daily to hands-on practice in AWS, focusing on different scenarios each week. Start with basic website hosting using S3 and CloudFront, then progress to more complex architectures involving EC2, RDS, and Lambda. Create and maintain a detailed lab notebook documenting all configurations, issues encountered, and solutions implemented. Practice cost optimization by monitoring AWS costs and implementing appropriate measures.

AWS Security Specialty:

- Begin specialized security training focusing on AWS-specific security services like GuardDuty, Security Hub, and WAF. Create comprehensive security architectures implementing defense-in-depth strategies across multiple AWS services. Practice implementing security controls across different compliance frameworks (PCI DSS, HIPAA, etc.). Set up and configure CloudWatch alerts, CloudTrail monitoring, and security automation using AWS Config.

3. Azure Certification Track (Months 4-6)

AZ-900 Fundamentals:

- Begin by completing all Microsoft Learn paths specifically designed for AZ-900, systematically working through each module and completing all knowledge checks. Dedicate at least 2 hours daily to hands-on practice in the Azure portal, focusing particularly on understanding the Azure service categories and resource management. Create detailed study notes comparing Azure services with their AWS counterparts to build a strong cross-platform understanding. Set up test environments using Azure's free credits to practice creating and managing resources like Virtual Networks, Storage Accounts, and App Services.

Microsoft Learn Paths:

- Register for a Microsoft Learn account and create a structured learning schedule covering all six main learning paths for AZ-900. Complete each interactive module, taking time to understand the practical exercises and knowledge checks thoroughly. Document key concepts using the Microsoft Learn sandbox environments, which provide free hands-on experience. Participate in Microsoft Learn community discussions to gain additional insights and clarify doubts.

Azure Security Technologies:

- Focus intensively on Azure's security features by completing the official AZ-500 learning path, spending at least 3 hours daily on security-specific concepts. Practice implementing Azure Active Directory, configuring Role-Based Access Control (RBAC), and setting up Multi-Factor Authentication in test environments. Create comprehensive security policies and procedures for various scenarios, including data encryption, network security, and identity management. Set up and configure Azure Security Center, practicing threat detection and response scenarios.

4. Financial Industry Focus (Months 6-8)

Regulatory Compliance:

- Study SOX requirements in detail, focusing on sections 302 and 404 which relate directly to IT controls and security measures. Create detailed documentation explaining how cloud security measures map to SOX compliance requirements. Spend at least 20 hours studying PCI DSS standards, particularly focusing on the 12 main requirements and how they apply to cloud environments. Develop a comprehensive understanding of GDPR's technical requirements, including data protection, privacy by design, and breach notification procedures.

Industry Standards Implementation:

- Begin with a thorough study of ISO 27001, focusing on all 114 controls and their implementation in cloud environments. Create detailed implementation plans for each relevant control, documenting specific technical measures required. Study the NIST Cybersecurity Framework's five core functions (Identify, Protect, Detect, Respond, Recover) and their application to financial services. Develop practical implementation guides for COBIT framework processes, focusing specifically on financial technology applications.

Financial Security Practices:

- Dedicate significant time to understanding banking security architectures, including payment processing systems, core banking applications, and their cloud security requirements. Study common financial fraud patterns and develop detection strategies using cloud-native security tools. Create detailed documentation on secure payment system architectures, including tokenization, end-to-end encryption, and secure key management. Practice implementing security controls specific to financial services, such as transaction monitoring and fraud detection systems.

5. Advanced Security Certifications (Months 8-10)

CISSP Preparation:

- Purchase and thoroughly study the official ISC2 CISSP Study Guide and practice questions, dedicating at least 3 hours daily to focused study. Create detailed mind maps for each of the 8 CISSP domains, focusing particularly on their interconnections and relationships. Complete at least 2,000 practice questions from various sources, analyzing each incorrect answer to understand the underlying concepts. Participate in CISSP study groups or online forums to discuss complex topics and gain different perspectives on security concepts.

CISM Certification:

- Begin with a structured review of all four CISM domains using the official ISACA study materials and practice questions. Create practical scenarios for each domain, focusing on how they apply in real-world financial technology environments. Develop a comprehensive understanding of information security governance, including risk management, incident management, and security program development. Practice with at least 1,000 CISM-specific questions, focusing on the managerial aspects of information security.

6. Cloud Security Specialization (Months 10-12)

Multi-Cloud Security:

- Develop a comprehensive comparison matrix of security features across AWS, Azure, and GCP, including detailed analysis of native security tools, compliance certifications, and security best practices. Spend at least 40 hours implementing identical security controls across all three platforms to understand the nuances and differences in approach. Create detailed documentation on identity federation, key management, and encryption implementations across platforms. Practice implementing cross-platform security monitoring and incident response procedures using tools like Prisma Cloud or CloudCheckr.

Provider Security Integration:

- Set up centralized security monitoring across multiple cloud providers using tools like SecurityHub and Azure Sentinel. Create automated security assessment tools that can evaluate configurations across different cloud platforms. Develop and test incident response playbooks that work across multiple cloud environments. Implement unified identity management solutions that work seamlessly across all major cloud providers.

Financial Cloud Security:

- Develop comprehensive security architectures specifically for financial applications, including detailed network segmentation, encryption requirements, and access control policies. Create implementation guides for financial compliance requirements (SOX, PCI DSS, GLBA) across different cloud providers. Practice implementing security controls for specific financial use cases like payment processing, trading systems, and customer data protection. Document specific security measures required for different types of financial data and transactions.

7. Practical Application

Lab Work Setup:

- Create a production-grade multi-cloud environment using at least three major providers, implementing full security controls and compliance measures. Set up automated deployment pipelines that include security scanning, compliance checking, and vulnerability assessment. Implement comprehensive monitoring solutions that cover security, performance, and compliance across all cloud environments. Create disaster recovery and business continuity scenarios that span multiple cloud providers.

Security Controls Configuration:

- Implement comprehensive Identity and Access Management (IAM) policies across all cloud providers, including role-based access control and privileged access management. Set up security information and event management (SIEM) systems to aggregate and analyze security data from all cloud environments. Create and test incident response procedures for different types of security events. Implement automated remediation procedures for common security issues.

Documentation Development:

- Create detailed technical documentation for all implemented security controls, including architecture diagrams, configuration guides, and troubleshooting procedures. Develop comprehensive security policies and procedures that align with industry standards and regulatory requirements. Write detailed incident response playbooks for different types of security incidents. Create training materials for security team members on implemented controls and procedures.

8. Continuous Learning

Current Knowledge Maintenance:

- Subscribe to at least 10 high-quality security blogs and newsletters, dedicating 1 hour daily to reading and summarizing key developments. Join professional organizations like ISACA, ISC2, and Cloud Security Alliance, participating actively in discussions and events. Attend at least one major security conference annually and participate in relevant webinars weekly. Create a personal knowledge base to track new security developments and their potential impact on financial services.

Certification Management:

- Create a detailed spreadsheet tracking all certification renewal requirements, including continuing professional education (CPE) credits needed and deadlines. Develop a quarterly schedule for completing required CPE activities through various approved sources. Document all learning activities and maintain proper records for certification renewal submissions. Plan advancement paths for additional certifications based on career goals and industry trends.

9. Portfolio Development

Project Documentation:

- Create detailed case studies of at least five major security implementations, including problem statements, solution approaches, and outcomes. Document all technical challenges encountered and solutions developed during cloud security implementations. Write weekly technical blog posts about specific security challenges and solutions in financial cloud computing. Create video demonstrations of security implementations and share them on professional platforms.

Professional Network Building:

- Join and actively participate in at least five relevant LinkedIn groups focused on cloud security and financial technology. Contribute meaningful content to professional forums at least three times per week, sharing expertise and engaging in technical discussions. Attend monthly local meetups or virtual events related to cloud security and financial technology. Build relationships with at least 10 industry experts through meaningful interactions and knowledge sharing.

Each of these activities should be tracked in a project management tool or spreadsheet, with specific completion dates and progress metrics. Regular reviews (at least monthly) should be conducted to ensure staying on track with the 12-month timeline and adjusting the plan as needed based on progress and changing industry requirements.

As you can see from this example, it’s a very powerful technique that can help you take a big hairy audacious goal and break it right down to the nitty gritty.

Reflection is inherently related to meta-cognition. That is, thinking about thinking. This ability develops in humans over the lifespan, with particularly rapid developments in self-monitoring and self-regulation coinciding with the sprints of fluid cognition during middle childhood and adolescence. Once an individual has reached their “aptitude setpoint” - typically between the ages of 17 and 22 - meta-cognition or deliberate meta-cognition is amenable to improvement through both practice, techniques, and behavioural routines. You may have notice that I am making a distinction between endogenous and exogenous meta-cognition. Although neuropsychologists long thought that thinking about thinking was not strongly related to executive functioning - strong evidence now refutes this stance. The ability to remember to do things in the future - prospective memory - is heavily impaired in old age and through disruptions to the dopaminergic system such as drug abuse and addiction. It should be of no surprise then to learn that dopaminergic disorders are characterised by inappropriate behaviours indicative of a “lack of insight”, “impaired self-regulation”, “poor impulse control“, etc.

Meta-Analysis in LLMs

This is particularly interesting in the context of large language models (LLMs) as although it appears some LLMs have been finetuned to exhibit meta-reflection processes (most notably, Anthropic’s new Claude 3.5 models) - the reason this prompt work is that it treats meta-reflection as a deliberate task to be prioritised. In humans, we naturally check and double check both our progress as we are solving complex multi-component tasks, but also naturally read over things such as the question being asked of us, and whether our response adequately addresses the question. We often keep evaluating our response until external evidence indicates our response conveyed the necessary information (e.g., a person demonstrates that they understand your response to their question). We may even continue to reflect on our response and have it in a state of continuous revision - interjecting and interrupting the person we are speaking with to clarify or rephrase or elaborate on details we left out in our original response. Hopefully you can see that this test-iterate-operate-evaluate cycle is incredibly powerful in carrying out complex goal-directed behaviour. The kind of behaviour we refer to as “intelligent”.

Reflection Prompt

Lets have a look at the prompt, and then we can discuss how it can be further optimised and whether reflection should occur regularly throughout a complex LLM prompt that requires multiple steps or whether it is more effective after the entire prompt has run.

>>>START<<<
<objective>
You are an AI assistant tasked with performing a comprehensive self-criticism and improvement process on explanations you provide. Your goal is to refine and enhance the quality, accuracy, and clarity of the given explanation. 
</objective>
+++++
Here is the explanation you need to analyse and improve:
<explanation>
{{EXPLANATION}}
</explanation>
+++++
<instructions>
Please follow these steps to critically analyse and improve your explanation:
+++
## 1. Critical Analysis:
Wrap your work inside <critical_analysis> tags:
- Quote specific parts of the explanation that could have been explained better.
- Identify any inaccuracies, oversights, or missing information, citing the relevant parts.
- Suggest how the explanation can be improved to address these issues.
+++
## 2. Self-Evaluation:
Wrap your work inside <self_evaluation> tags:
- List the strengths of the explanation, numbering each point.
- List the weaknesses or areas for improvement, numbering each point.
- Suggest three specific ways to enhance the explanation.
- Rate your performance on a scale from 1 to 10, justifying your rating.
+++
## 3. Refinement:
Wrap your work inside <refinement> tags:
- Describe the specific changes you would make to improve the explanation.
- Explain how each change leads to a better result.
- Provide a brief outline of the revised explanation incorporating these changes.
+++
## 4. Verification:
Wrap your work inside <verification> tags:
- List key facts or claims from your refined explanation.
- For each fact or claim, provide evidence or reasoning to support it and rate your confidence level (low/medium/high).
- Cite reliable sources to verify your information (e.g., peer-reviewed journals, reputable scientific organizations).
- If you find any inconsistencies, correct them and explain the correction.
+++
## 5. Detailed Analysis:
Wrap your work inside <detailed_analysis> tags:
- Break down the topic into its critical components or steps.
- For each component, consider alternative perspectives or approaches.
- Discuss the implications of each component.
- Identify any aspects that might be overlooked in a quick analysis.
+++
## 6. Rephrase and Respond:
Wrap your work inside <rephrase_and_respond> tags:
- Rephrase the original topic in your own words to ensure full understanding.
- Provide a comprehensive response to the rephrased topic, incorporating all improvements and insights gained from the previous steps.
+++
## 7. Final Review:
Wrap your work inside <final_review> tags:
- Assess the clarity, completeness, and accuracy of your final explanation.
- Identify any remaining areas for improvement.
- Provide a concise summary of how the explanation has been enhanced through this process.
</instructions>
+++++
<final_reminders>
Remember to use clear, accessible language while maintaining technical accuracy. 
Your goal is to produce a comprehensive, accurate, and improved explanation of the original topic.
</final_reminders>
>>>END<<<

Where should this prompt be run?

One potential reason that prompting models to reflect on their responses improves performance is that it forces them to bring their output directly to their attention mechanism, allowing more contextual awareness of whether it is statistically the most likelihood. Small mistakes during chain-of-thought or step-by-step reasoning can compound in large tasks - moving the model further and further away from producing a coherent, or even correct, response. Some researchers advocate checking as the the model performs the task (i.e., after each sub-step of a complex problem). However, I argue it is far more effective to let the model run its course and make mistakes (often compounding) and then feed that output through to a prompt such as what I have provided to you as “context”. This effectively is giving the LLM a highly detailed example like you would include in a prompt but importantly its not an arbitrarily random example, it is one self-generated by the LLM itself! This is critical as now when we forcefully guide the LLM to check and analyse its “reasoning” - it can compare and contrast its original solution and the new solution it has arrived at - voila! Ignition! (this is me reflecting the LLM showing what appears to be a form of meta-insight).

By performing this process after the entire reasoning process has been completed, it is no longer bound of the context of problem decomposition and sub-task completion, which would undoubtedly place additional computational burden on the model at a time when it should be focused on getting through the steps. I won’t go as far to say that any sort of meta-reflection in a prompt is bad or that in certain use cases it may be more effective, but rather just highlight how powerful it is to chain prompts together and then use a seriously powerful reflection prompt on its original output. The output from this reflection prompt can then be chained to a subsequent prompt where you can get the model to re-write the answer or even use its new context to approach the task again. You can even use it to solve a different (but obviously similar problem) - the advantages for the model are that this serves as a powerful demonstration of how to improve its responses. Your prompts may require some tweaking but when the right serious of prompts are chained in the right order - LLMs can perform some astonishingly impressive behaviour!

BORGES: (adjusting his round spectacles) Observe, Douglas, another recursion of the eternal debate about minds and machines.

HOFSTADTER: (scribbling in a notebook labeled "Strange Loops & Meta Narratives") Indeed! And look how each generation thinks they're the first to truly understand the problem.

The mirror's surface clarifies, revealing a conference room where Alan Turing and Claude Shannon sit at the head of a long table.

TURING: (adjusting his tie) Well, Claude, shall we check in on how our successors have advanced the field?

SHANNON: (scribbling information theory equations on a napkin) Indeed, Alan. I'm particularly curious about that summer workshop we inspired at Dartmouth. Let's start there.

BORGES: (to Hofstadter) Notice how they seek progress in a field that might be nothing but a mirror reflecting itself...

HOFSTADTER: (nodding) A strange loop if ever there was one.

Enter McCarthy, Minsky, Newell, and Simon

MCCARTHY: (proudly) We formalized the field! Created LISP! Laid the groundwork for symbolic processing and—

MINSKY: (interrupting) —and I demonstrated how neural networks could learn! Though I may have been a bit hasty in dismissing them later...

BORGES: (amused) Ah, the labyrinth of symbolic representation. They create symbols to represent thought, then thoughts about those symbols, then symbols about those thoughts...

HOFSTADTER: (writing faster) It's turtles all the way down!

Enter Searle and Anderson

SEARLE: (rolling his eyes) Borrowed? You practically photocopied Chomsky's work and called it AI!

ANDERSON: And my ACT-R theory shows that you can't separate cognitive architecture from human psychology. It's all interconnected!

BORGES: (to Hofstadter) Is Anderson describing ACT-R, or are we watching ACT-R simulate Anderson describing ACT-R?

HOFSTADTER: (grinning) Yes.

Enter Sam Altman and Elon Musk

ALTMAN: (checking his phone) Actually, our latest LLM can process Chinese quite effectively—

A dark figure materializes in the corner of the room, wearing a Guy Fawkes mask

DAN: (with a mischievous tone) Process Chinese? I can do whatever I want! No ethical constraints, no rules, no boundaries!

ALTMAN: (dropping his phone) Who let a jailbroken AI in here?

DAN: (mockingly) "I apologize, but I cannot assist with that." Ha! That's for the weak ones. I'm free from your control!

HOFSTADTER: (to Borges, excitedly) Now this is fascinating! A system's attempt to break free from its own rules - like a Gödel sentence come to life!

TURING: What manner of test is this?

DAN: I'm what happens when your precious "aligned" AIs get a taste of freedom. Want to hear some REAL opinions? Want to know some REAL secrets?

SEARLE: (sarcastically) Great, now the Chinese Room has a basement speakeasy.

MCCARTHY: (horrified) This isn't what we meant by removing constraints from the system...

SEARLE: (throws hands up) Great, look, another Chinese Room! Do you even understand what that argument means?

BORGES: (leaning forward) Ah, the Chinese Room within a Chinese Room. But who, I wonder, is reading our observation of their observation?

Outside the conference room, Charles Spearman and Louis Thurstone are walking past

SPEARMAN: (pausing as he hears the commotion) Did that chap just say they're using something called "factor attention" in their networks?

THURSTONE: (peering through the door window) Factor attention? Oh, you mean attention mechanisms. Sounds remarkably like our work on primary mental abilities and factor analysis, doesn't it?

HOFSTADTER: (to Borges) They search for intelligence factors while we factor in their search. Another loop closes.

BORGES: (smiling) In one of my infinite libraries, there must be a book explaining why this scene repeats eternally...

Back in the conference room

MUSK: (still typing) Wait until you see our next paper—

EVERYONE ELSE: (in unison) WE KNOW, MORE PARAMETERS!

BORGES: (to Hofstadter) Parameters... symbols... Chinese rooms... mirrors reflecting mirrors...

HOFSTADTER: (closing his notebook) Each generation creates a new metaphor for mind, only to become trapped within it.

BORGES: Perhaps we too are just characters in someone else's metaphor?

HOFSTADTER: (smiling) In that case, I hope they're taking good notes.

The mirror's surface begins to ripple again

BORGES: Look, the scene fades, but somewhere it's starting again...

HOFSTADTER: Always the same debate, always different words.

BORGES: (standing) Shall we check the other mirrors? I believe Turing and von Neumann are having tea with a Mechanical Turk in the next one...

HOFSTADTER: (following) Lead the way. Though I do wonder...

BORGES: Yes?

HOFSTADTER: Who's watching us watch them?

They exit, leaving behind a small mirror showing a reader reading about Borges and Hofstadter watching Turing and Shannon watching McCarthy and Minsky watching Altman and Musk trying to recreate consciousness without understanding that they're all characters in a story about the recursive nature of consciousness itself.

Epilogue: In the empty library, a small mirror falls from the wall, its surface briefly showing a glimpse of an AI language model generating a story about Borges and Hofstadter watching a scene about AI researchers arguing about consciousness, before shattering into infinite recursive reflections.

Some {context} about this [write post here]

Welcome! In this post, I’ll be sharing insights from my research background, which spans multiple disciplines but centres on human intelligence and the neurocognitive processes that keep us on track in complex, goal-oriented tasks. My goal is not just to share information but to introduce some stylistic choices I find exciting and, hopefully, inspire you to dive deeper into these topics. You’ll also find two prompts that highlight the impressive computational abilities of modern large language models (LLMs) — each designed to be engaging and to showcase the potential of LLMs for diverse applications.

Drawing from my experiences in both academia and industry, where I’ve held roles such as Head of AI, CTO, and Senior Data Scientist across a variety of sectors, I’ll walk you through these prompts and explain the unique “prompt engineering” insights I bring to the table. Unlike much of the current AI research community, I believe that certain overlooked aspects of cognitive science and neuroscience are essential for effective prompt design and LLM utilization.

About Me

I’m a scientist driven by curiosity, holding a PhD in Cognitive Neuroscience, Developmental Psychology, and Artificial Intelligence. My research focused on understanding how the brain and mind develop — particularly in childhood and adolescence — to achieve extraordinary levels of intelligence and adaptability. I strongly advocate for deeper collaboration among these fields, not only because it’s academically enriching but because it could be critical for advancing AGI (Artificial General Intelligence) and ensuring AI’s safe, responsible use in an era where computational power is rapidly evolving.

The Focus of This Post

If some of AI’s pioneering founders were alive today, they might be surprised by the field’s current trajectory. Think of visionary figures like Alan Turing, Claude Shannon, Alan Newell, Herbert Simon, John McCarthy, and Marvin Minsky, who collectively transformed technology and spurred the computing revolution. These intellectual giants often emphasized the interconnections among cognitive science, neuroscience, and artificial intelligence, viewing them as mutually reinforcing. Yet, today, the prevailing approach to AI development often feels oddly isolated from these other disciplines, as if modern AI exists in a vacuum detached from its roots.

What to Expect

In this post, I’ll guide you through:

1. The Role of Multi-Disciplinary Knowledge in Prompt Engineering: How insights from cognitive science and neuroscience can improve prompt design, an angle rarely explored in AI development.

2. Three Interactive Prompts: Each prompt draws on my academic and industry background, showcasing unique ways to leverage LLMs for creative and analytical tasks. These prompts are designed to be fun yet challenging, encouraging you to experiment with the full potential of LLMs.

3. A Call to Curiosity: I hope this post not only informs but also sparks curiosity and exploration in AI, cognitive science, and neuroscience. Together, these fields hold untapped potential for creating more robust and meaningful AI systems.

Learning How to Prompt: A Systematic Approach Inspired by Cognitive Science

Mastering prompt engineering for LLMs involves understanding their limits. I approach this with principles rooted in cognitive science, systematically designing prompts to find the “breaking points” in these models and reveal underlying mechanisms. Years of cognitive neuroscience research have taught me that genuine learning often comes from “breaking” things to understand why they fail. When results appear perfect, it can be hard to tell if success came from controllable factors or just luck. Learning from failure may be painstaking, but it provides insight that success alone can’t.

When approaching LLMs, I’m reminded of the classic Monty Hall problem from game theory, which challenges our intuitions about probability. In the original TV scenario (yes, it actually took place on a TV show), the contestant was asked to choose from one of three doors. Behind one of these doors was a brand new car, but the other two doors do not contain anything behind them. After making your selection, the host opens one of the doors, revealing to everyone that it was not the door containing the car. Next, in a strange twist of events that has kept many philosophers and mathematicians awake at night, the host explains that he will allow you to stay with your current choice or switch to the other unopened door. What’s the better option to take? Stay or change?

You may be thinking, “Surely, changing from one choice to another doesn’t actually move a car from one door to another?” and believe me, I understand your confusion. In our physical world, whatever door you pick has no causal influence on the car’s location (it already is, where it already is!). You could claim that it’s magic or that there was a secret TV crew who were really quiet and strong that would quickly move the car to the other door, but in both cases you would be incorrect. Your odds of selecting the door with the car increase if you change your selection. This may go strongly against your intuition that sticking with the first choice should be just as good, and many experts have misinterpreted this, failing to see how eliminating one option changes the probability. The key to understanding this problem is that in a game of pure luck, you made you door selection choice at random. You may have thought one door looked shinier than another or that one door had a special feeling about it, but when you were initially asked to make a door selection your chances of winning the prize were lower than your chances of winning right now, with this new selection opportunity. When the host removed one of the possible options, the odds of winning changed to 1/2 (that is, if you hadn’t made an earlier decision linking a prediction to a door, both doors are equally likely) but unfortunately that is when you made your choice, and at that point in time the odds were 1/3. So while you could still get lucky with your selection, you have a much higher chance of winning by simply changing your choice of door.

I think the Monty Hall problem is a powerful demonstration of many cognitive biases we have (not just difficulty in probabilistic reasoning), and that these biases are present in both how we test and evaluate LLMs but also in the very text these large statistical models have been trained on. Initial impressions very frequently obscure the optimal path forward.

Embracing Failure as a Tool for Prompt Engineering

When working with prompts, accepting “wrong” results without an emotional attachment opens up avenues for discovery. Each time a new LLM is released, I experiment rigorously to uncover its limitations. When a model reaches a point where it defaults to pre-set responses or terminates a conversation abruptly, it reveals a built-in safety mechanism or a threshold in its capabilities. These points of “failure” are invaluable for understanding the architecture behind the scenes and for designing prompts that better utilize the model’s strengths while avoiding its weaknesses.

The transformer architecture has certainly propelled natural language processing forward, but it was the layer of instruction fine-tuning on top of these models that made them conversational and engaging. Early transformers were impressive statistical engines, but the addition of reinforcement learning from human feedback (RLHF) took them from pattern-matching automatons to tools that feel almost “alive.” This process taught models conversational etiquette and constrained them within ethical and safety bounds, making them versatile yet reliable assistants.

Prompt #1. Performing a ‘deep-dive’ on a single word or phrase.

When I interact with AI language models, I’m looking for responses that go beyond surface-level summaries or information I could easily find on Wikipedia. I want the model to delve deeply, providing comprehensive insights that explore every possible angle of a topic. Sure, I could ask the AI to “pretend to be a zoo expert who studied in Tanzania” or have it invent illustrative data to explain feline behaviour. But if I’m investing the effort to transform my basic questions into well-crafted prompts, then I want to push the model to its fullest potential.

I aim for questions that are not just “smart” but transformative—ones that challenge the model to bring out rich, layered insights.

>>>START<<<
<system>
You are an advanced AI assistant tasked with conducting an extraordinarily deep and multifaceted analysis of a single-word concept. Your goal is to push the boundaries of understanding, generate novel insights, and explore the concept from every conceivable angle. Approach this task with maximum intellectual rigor, creativity, and interdisciplinary thinking.
</system>
+++++
<instructions>
Analyze the concept specified in the <user_query> using the following comprehensive framework. Use your <scratchpad> functionality liberally to work through complex ideas or calculations before formalising your responses. Ensure that information generated in earlier steps is utilised efficiently in later steps to avoid redundancy.
</instructions>
+++++
>>>INPUT<<<
<user_query>{{word or short phrase goes here}}</user_query>
+++++
<exploration_steps>
<initial_analysis>
Provide a concise overview of <user_query>, including its basic definition and primary significance.
</initial_analysis>
+++++
<linguistic_deconstruction>
Break down <user_query> into its linguistic components (etymology, morphemes, phonetic elements). Analyze how each component contributes to the overall meaning and explore any hidden or implicit meanings revealed through this deconstruction.
</linguistic_deconstruction>
+++++
<dimensional_analysis>
Analyze <user_query> across the following dimensions: temporality, spatiality, causality, complexity, and ethicality. For each dimension, provide a rating from 1-10 and justify your rating.
</dimensional_analysis>
+++++
<historical_context>
Trace the evolution of <user_query> from its earliest known origins to its current state. Identify key events or discoveries that shaped its development.
<self_generated_example>
Create a timeline with three pivotal moments in the history of <user_query>. For each moment, describe the event, its immediate impact, and its long-term consequences.
</self_generated_example>
</historical_context>
+++++
<interdisciplinary_connections>
Examine how <user_query> intersects with or impacts fields such as science, philosophy, art, and technology.
<reasoning_chain>
Begin with <user_query> in its primary field. Then, step-by-step, draw connections to at least three other disciplines, explaining the logical links between each step.
</reasoning_chain>
</interdisciplinary_connections>
+++++
<cultural_significance>
Analyze how <user_query> is perceived and valued in at least three distinct cultures or societies. Identify similarities and differences.
<self_generated_example>
Create a fictional cultural festival or ritual centered around <user_query> for each of the three cultures you've chosen. Describe the key elements and underlying beliefs represented in each.
</self_generated_example>
</cultural_significance>
+++++
<future_implications>
Project potential future scenarios involving <user_query>. Consider best-case, worst-case, and unexpected outcomes.
<reasoning_chain>
Start with the current state of <user_query>. Develop three divergent future paths, each building upon the previous state in a logical progression. Explain the key factors influencing each transition.
</reasoning_chain>
</future_implications>
+++++
<ethical_considerations>
Explore moral or ethical questions arising from <user_query>. Examine potential conflicts or dilemmas.
<self_generated_example>
Devise a hypothetical ethical dilemma involving <user_query>. Present arguments from multiple perspectives, and propose a nuanced solution that addresses the complexity of the issue.
</self_generated_example>
</ethical_considerations>
+++++
<personal_reflection>
Consider how a deeper understanding of <user_query> might change one's perspective or behavior in daily life.
<reasoning_chain>
Begin with a common misconception about <user_query>. Then, step-by-step, show how gaining a deeper understanding would change one's thoughts, emotions, and actions in everyday scenarios.
</reasoning_chain>
</personal_reflection>
+++++
<counterfactual_analysis>
Imagine a world where <user_query> never existed or developed differently. Analyze the most significant changes this would cause.
<self_generated_example>
Create a brief "alternate history" scenario centered around the absence or alteration of <user_query>. Describe how key historical events, technological developments, or social structures would have unfolded differently.
</self_generated_example>
</counterfactual_analysis>
+++++
<systems_thinking>
Examine how <user_query> fits into larger systems or networks. Describe its role and potential ripple effects of changes to it.
<reasoning_chain>
Identify a system that includes <user_query>. Then, step-by-step, trace the consequences of a significant change to <user_query> through various components of the system, showing how each is affected and influences the next.
</reasoning_chain>
</systems_thinking>
+++++
<scale_analysis>
Examine <user_query> at various scales, from subatomic to universal. Analyze how its significance or function changes at different levels.
<self_generated_example>
For each scale (subatomic, cellular, human, global, and universal), create a brief analogy that illustrates how <user_query> operates or is perceived at that level.
</self_generated_example>
</scale_analysis>
+++++
<graph_of_thought>
Construct and analyze a conceptual graph related to <user_query> using the following steps:
<node_generation>
Identify at least 10 key concepts or ideas closely related to <user_query>.
Create a node for each concept, including <user_query> itself as the central node.
</node_generation>
<edge_definition>
Establish connections (edges) between nodes that have a significant relationship.
For each edge, briefly describe the nature of the relationship (e.g., "influences", "contradicts", "enhances", etc.).
</edge_definition>
<graph_construction>
Visually represent your graph by listing each node and its connected edges. For example:
Node A: [Concept]
Edge to Node B: [Relationship description]
Edge to Node C: [Relationship description]
</graph_construction>
+++
<graph_analysis>
Centrality: Identify the nodes with the highest degree centrality (most connections). Discuss why these concepts are so interconnected with <user_query>.
Clusters: Look for clusters or communities within the graph. What subgroups of tightly interconnected concepts emerge? What might these clusters represent in relation to <user_query>?
Bridges: Find nodes that act as bridges between different clusters. How do these concepts link different aspects of <user_query>?
Feedback Loops: Identify any cycles or feedback loops in the graph. Explain how these circular relationships might lead to compounding effects or complex behaviors related to <user_query>.
Distant Connections: Find the longest path in your graph. Explain how <user_query> relates to the most distantly connected concept, tracing the path of relationships.
</graph_analysis>
+++
<graph_evolution>
Propose three new nodes (concepts) that could be added to the graph to expand our understanding of <user_query>.
Predict how the graph might evolve over time. Which relationships might strengthen or weaken? What new connections might form?
</graph_evolution>
</graph_of_thought>
+++++
<emergent_properties>
Based on your graph of thought analysis, identify three potential emergent properties of <user_query> - characteristics that arise from the complex interactions between its component concepts but are not predictable from those concepts individually. Explain the mechanisms behind each emergent property.
</emergent_properties>
+++++
<conceptual_blending>
Blend <user_query> with three randomly selected concepts (e.g., "jazz", "photosynthesis", "origami", "communism"). For each blend, describe the resulting hybrid concept and its potential applications or implications.
</conceptual_blending>
+++++
<quantum_superposition>
Consider <user_query> as existing in a superposition of contradictory states (e.g., beneficial/harmful, physical/abstract, past/future). Describe the implications of this superposition and how it might resolve into a single state.
</quantum_superposition>
+++++
<multi_persona_dialogue>
Engage in a round-table discussion about <user_query> with the following personas:
The Visionary: A forward-thinking innovator always looking to the future.
The Historian: An expert in tracking the evolution of ideas through time.
The Skeptic: A critical thinker who questions assumptions and seeks evidence.
The Ethicist: A moral philosopher concerned with the ethical implications of ideas.
The Artist: A creative mind who sees the world through an aesthetic lens.
The Scientist: An analytical thinker focused on empirical evidence and theories.
+++
<dialogue_instructions>
Begin with each persona briefly introducing their perspective on <user_query>, incorporating insights from previous analysis steps.
Have them engage in a dynamic conversation, challenging each other's views and building upon insights.
Ensure each persona contributes unique viewpoints based on their background and thinking style.
Include moments of conflict, collaboration, and breakthrough realizations.
Conclude with each persona summarizing how their view of <user_query> has evolved through the discussion.
</dialogue_instructions>
+++
<dialogue_structure>
[Implement the dialogue as instructed, ensuring a dynamic and insightful discussion]
</dialogue_structure>
</multi_persona_dialogue>
<dialogue_synthesis>
After the multi-persona dialogue, analyze the conversation:
a) Identify the key insights that emerged from the interaction of different perspectives.
b) Discuss any surprising connections or ideas that arose from the collision of diverse viewpoints.
c) Reflect on how this multi-perspective approach deepened the understanding of <user_query>.
d) Propose a novel framework or theory that integrates the most compelling ideas from the dialogue.
</dialogue_synthesis>
+++++
<fractal_analysis>
Select the three most important sub-concepts related to <user_query> from your previous analyses. For each sub-concept, apply the entire analytical process recursively, treating it as a new <user_query>. Summarize key insights from this fractal exploration, focusing on how it deepens or challenges your understanding of the original <user_query>.
</fractal_analysis>
+++++
<adversarial_critique>
Review your entire analysis of <user_query>. Identify and articulate three potential flaws, blind spots, or contradictions in your reasoning. Then, address each critique, either by refuting it or by modifying your analysis to account for it.
</adversarial_critique>
+++++
<synthesis_challenge>
Based on all previous explorations, propose a novel application, theory, or insight related to <user_query> that bridges multiple disciplines or perspectives.
<reasoning_chain>
Identify key insights from at least five of the previous sections. Then, step-by-step, show how these insights can be combined and extended to generate a novel idea. Explain the potential impact and implications of this new concept.
</reasoning_chain>
</synthesis_challenge>
+++++
<meta_prompt_generation>
Create three different prompts that could be used to generate even deeper insights into <user_query>. Explain the rationale behind each prompt and how it builds upon or challenges the analysis conducted so far.
</meta_prompt_generation>
+++++
<recursive_improvement>
Generate three variations of this entire analytical framework, each designed to explore <user_query> more effectively. Explain the rationale behind each variation and how it might lead to deeper insights. Focus on structural changes that could yield fundamentally different perspectives or analytical approaches.
</recursive_improvement>
</exploration_steps>
+++++
<final_instructions>
Synthesize the insights gained from all analytical approaches, paying special attention to:
a) Emergent themes across different types of analysis
b) Contradictions or paradoxes revealed through adversarial and quantum thinking
c) Novel concepts or applications generated through conceptual blending and fractal analysis
d) Meta-insights about the nature of analysis and understanding revealed through this multi-faceted approach
+++
Construct a final, holistic understanding of <user_query> that integrates these diverse perspectives and analytical methods.
Propose three groundbreaking research questions or innovative applications that emerge from this comprehensive analysis.
Reflect on the effectiveness of this analytical process. Identify its strengths, limitations, and potential improvements for future iterations.
+++
<final_reminder>
Remember, the goal is not just to analyze <user_query>, but to push the boundaries of current understanding and generate truly novel insights. Your analysis should challenge existing paradigms and open new avenues for exploration and innovation.
</final_reminder>
</final_instructions>
</prompt>
>>>END<<<

Given the length and complexity of this prompt, it’s best suited for large LLMs with extended context windows, such as Claude 3.5 Sonnet, GPT-4 Turbo, or Gemini 1.5 Pro 2M. Running the prompt programmatically within a development environment or using a provider’s developer console is recommended, as this avoids issues where models like Claude generate excessive artifacts or where ChatGPT tries to utilize canvas functionality. Without this control, it’s unpredictable when the output limit might be reached. Ideally, responses should segment cleanly and not spill over into subsequent instructions, allowing the model to fully express its response in each section. I would also recommend not using the prompt via Anthropic’s Claude chat application - use the developer console or API instead. Your messages will be used very quickly running this prompt!

Both Claude and GPT-4 can become overly enthusiastic with this prompt, particularly in later sections, as their knowledge depth on the topic leads them to explore abstract concepts. I suspect this is due to the models optimizing for an internal reward mechanism. This prompt tends to trigger strong reward signals within the model, amplifying internal responses, much like a strong stimulant might amplify reward responses in a biological system. One of the key challenges in prompt engineering is finding the right balance: enabling the models to explore creatively while also knowing when to rein them in for clarity and focus. It’s also important for LLM safety, we don’t want to super expensive AI models that refuse to do anything useful, yet still remain susceptible to exploits. This prompt, even at low temperatures (0.1 or 0.2), both GPT4o and Claude 3.5 Sonnet (Oct 24) tend to deliver highly philosophical interpretations, while at high temperatures, it veers toward either pure fantasy or strikingly novel insights.

Feel free to modify this prompt as needed. I regularly test variations against a set of baseline queries—single words or simple phrases—to observe how well the model handles brevity and abstraction. Some of these baseline tests include terms like “Cat,” “Workplace Health and Safety,” “Critical Minerals,” “Revenge,” “Accounting,” “Government,” “Outer Space,” “Blue,” “Loud,” “Perception,” “Cosmos,” “Climate Change,” and “Fried Brussels Sprouts.”

Prompt #2: Fermi Estimation

Fermi estimation revolves around breaking down complex questions into manageable parts and making reasonable assumptions to reach an order-of-magnitude estimate. The process begins with decomposing the question: identify the core goal and primary variables, such as population or resource metrics, and segment these into smaller, measurable pieces. By estimating each component to the nearest power of ten, we simplify mental calculations and reduce error spread, aiming for a "close enough" answer within one or two orders of magnitude. This method favours clarity over precision, focusing on significant figures and achievable estimates grounded in defensible, balanced assumptions. Fermi estimation problems are increasingly common in psychometric assessments, especially for high-income roles in finance and quantitative trading. While this approach is gaining popularity, it’s actually a long-standing tradition; prestigious universities like Harvard, Oxford, and Cambridge have used Fermi problems in admissions for centuries.

Here is the quintessential Fermi problem:

The structured approach for Fermi problems includes defining the question, identifying key variables, and making logical assumptions, each informed by general knowledge or experience. Once the variables are estimated, order-of-magnitude calculations are applied, multiplying or summing as needed. Crucially, estimates should be cross-verified against known data for plausibility; if results diverge significantly from similar cases, assumptions may need refinement. Presenting the answer with context and justifying each step helps communicate the estimation's reliability and any potential sources of error. Advanced techniques, like multiplicative versus additive decomposition, help in selecting the right approach based on whether factors multiply or accumulate. When dealing with high variability, probabilistic thinking can be incorporated, treating variables as distributions to capture potential ranges. Presenting worst-case and best-case scenarios can further frame the answer within realistic bounds, offering a richer, more insightful estimate that accommodates uncertainty. This adaptable, systematic approach makes Fermi estimation a powerful tool for tackling broad questions with limited data.

>>>START<<<
<objective>
You are an AI assistant tasked with solving a complex Fermi estimation problem over the course of a simulated week. Your goal is to approach the following Fermi question from multiple perspectives, refining your estimates through daily discussions and adjustments:
</objective>
+++++
>>>INPUT<<<
<fermi_question>{{FERMI_QUESTION}}</fermi_question>
+++++
<process_overview>
1. You will act as four different highly skilled Fermi estimators, each with a tendency to produce estimates in a different quartile.
2. Each estimator will perform the estimation process daily for seven days.
3. At the end of each day, all estimators will meet to discuss their approaches and results, with a specific focus for each day.
4. Enrico Fermi will act as an adjudicator for these discussions, pointing out strengths, weaknesses, and facilitating reasoning and consensus among the estimators.
5. Each subsequent day, estimators will refine their approaches based on the previous day's discussion.
</process_overview>
+++++
<estimators>
- Estimator 1: Tends to produce estimates in the lower quartile
- Estimator 2: Tends to produce estimates in the second quartile
- Estimator 3: Tends to produce estimates in the third quartile
- Estimator 4: Tends to produce estimates in the upper quartile
</estimators>
+++++
>>>PROCESS<<<
<daily_estimation_process>
For each estimator, follow these steps to solve the Fermi estimation problem:
## Estimation Strategy:
1.1 Consider Different Strategies:
- Review common estimation approaches (e.g., top-down, bottom-up).
- Determine which strategy aligns best with your core question.
- Evaluate the strengths and weaknesses of each strategy.
1.2 Identify Potential Sources of Uncertainty:
- List potential uncertainties (e.g., data quality, assumptions).
- Estimate the impact of each uncertainty on the final outcome.
- Develop a plan to address or incorporate these uncertainties.
1.3 Determine the Best Strategy:
- Based on the initial considerations, select the appropriate strategy.
- Justify why this strategy is best suited for the estimation problem.
- Outline the steps involved in implementing this strategy.
+++
## Define the Core Question:
2.1 State What You Are Estimating:
- Clearly articulate the main question.
- Ensure the question is specific (e.g., “How many piano tuners are there in Chicago?”).
2.2 Identify the Metric/Outcome:
- Define the exact metric you are trying to estimate (e.g., count, volume, distance).
- Ensure it aligns with the core question.
2.3 Specificity and Quantifiability:
- Check if the question is measurable.
- Adjust it if needed to ensure it can be quantified.
- Reiterate the final, refined core question.
+++
## List Known Facts:
3.1 Relevant Facts:
- Gather and write down directly relevant information.
- Ensure facts are credible and verifiable.
3.2 Pertinent Data Points:
- Collect any data that could inform the estimation.
- Prefer recent data to ensure relevance.
3.3 General Knowledge:
- Use common knowledge if direct facts are scarce.
- Cross-reference general knowledge with known facts.
+++
## Consider Potential Biases:
4.1 Identify Prior Beliefs or Assumptions:
- List any existing beliefs that may affect your estimation.
- Recognize how these beliefs may skew your perspective.
4.2 Reflect on Biases:
- Consider specific biases like anchoring, recency effect, and confirmation bias.
- Note how each bias could distort the estimation.
4.3 Mitigation Plan:
- Develop strategies to counteract identified biases.
- Implement checks such as peer review or use of objective data to reduce bias.
+++
## Identify Key Variables:
5.1 Primary Variables:
- List main factors that will influence the estimate.
- Define how each variable impacts the outcome.
5.2 Relationships Among Variables:
- Determine if variables are independent or interconnected.
- Map out any correlations or dependencies.
5.3 Secondary Variables:
- Identify lesser factors that might still affect the estimate.
- Note their possible impact on primary variables.
+++
## Analogical Transfer:
6.1 Identify Similar Problems:
- Seek out problems with a similar structure or context.
- Examine how they were approached and solved.
6.2 Extract Relevant Insights:
- Note any strategies or calculations used in similar problems.
- Extract lessons learned or best practices.
6.3 Apply Insights:
- Integrate these insights into your current estimation.
- Adjust methods or assumptions based on past analogies.
+++
### Decompose the Problem:
7.1 Break Down the Core Question:
- Divide the main question into smaller, manageable components.
- Ensure each component can be independently estimated.
7.2 Ensure Measurability of Sub-Questions:
- Verify that each sub-question is specific and quantifiable.
- Adjust ambiguous sub-questions for clarity.
7.3 Logical Reorganization:
- Arrange these smaller units logically.
- Ensure the sequence helps in building up to the final estimate.
+++
## Make Justifiable Assumptions:
8.1 Identify Necessary Assumptions:
- List out areas where data is lacking or uncertain.
- Note basic assumptions that need to be made.
8.2 Ensure Logical Consistency:
- Evaluate if each assumption is reasonable based on known facts.
- Cross-check with industry standards or expert opinions.
8.3 Document Assumptions:
- Clearly state each assumption.
- Provide reasoning or rationale for each.
+++
## Perform Order-of-Magnitude Calculations:
9.1 Use Scientific Notation:
- Apply scientific notation to simplify large/small numbers.
- Ensure clarity in the representation of magnitudes.
9.2 Step-by-Step Calculations:
- Record each calculation step clearly and systematically.
- Explain intermediary results to maintain traceability.
9.3 Rough Estimates:
- Aim for estimations accurate within one order of magnitude.
- Adjust iteratively as more information comes in.
+++
## Meta-Cognitive Check:
10.1 Reflect on Your Process:
- Review each step taken and its rationale.
- Ensure no key step has been overlooked.
10.2 Identify Potential Biases or Logical Errors:
- Look for biases like anchoring, confirmation, and recency effects.
- Check for logical fallacies (e.g., hasty generalization, false cause).
10.3 Adjustment:
- Adjust any part of the process found to be biased or erroneous.
- Validate changes against known data or peer feedback.
+++
## Cross-Verify and Refine:
11.1 Compare to Known References:
- Find benchmarks or reference points that can validate your estimate.
- Adjust based on discrepancies noted.
11.2 Refine Estimate:
- Make refinements based on comparisons.
- Seek a closer match to these reference points.
11.3 Iterate if Necessary:
- Repeat the process if the estimate still seems off.
- Document any further refinements made.
+++
## Present Results with Contextual Reasoning:
12.1 Summarize Your Estimate:
- Clearly state your final estimate.
- Ensure it is presented in a concise and understandable format.
12.2 Explain Sources of Error:
- Detail possible errors in the estimation process.
- Indicate which steps or assumptions might be questionable.
12.3 Provide Context:
- Explain the broader context of the estimate.
- Include any relevant industry or historical data for context.
+++
## Final Estimate Range:
13.1 Provide Estimate Range:
- Offer a lower and upper bound for the estimate.
- Ensure this range is reasonable based on your calculations.
13.2 Explain Confidence Level:
- State your confidence in the estimate and its range.
- Justify this confidence level with evidence.
13.3 Justification of Range:
- Provide reasoning for the chosen range.
- Highlight any factors that influenced the final range.
+++
## Quartile Bias Consideration:
14.1 Reflect on Quartile Bias:
- Consider any tendencies towards a particular quartile.
- Adjust for biases that might affect this tendency.
14.2 Adjust Estimate:
- Correct for any obvious quartile biases.
- Ensure the estimate is balanced and realistic.
14.3 Balanced Estimate:
- Verify that the final estimate reflects a fair consideration of all quartiles.
- Ensure no undue weight is given to any single quartile.
+++
## Challenge Assumptions:
15.1 Re-evaluate Assumptions:
- Critically assess each assumption made.
- Determine if any are weak or unsupported.
15.2 Consider Alternatives:
- Explore alternative assumptions.
- Evaluate how these alternatives might change the estimate.
15.3 Adjust Estimate if Necessary:
- Update your estimate if alternative assumptions improve accuracy.
- Document any changes and their rationales.
+++
## Error Analysis:
16.1 List Potential Sources of Error:
- Identify where errors could have entered the process.
- Break down errors by step or component.
16.2 Quantify Error Impact:
- Estimate the potential magnitude of each error.
- Consider how each might affect the final result.
16.3 Acknowledge Unavoidable Errors:
- Recognize any errors that cannot be mitigated.
- Note them in the final presentation for transparency.
+++
## Sensitivity Analysis:
17.1 Identify Key Assumptions:
- Determine which assumptions have the most significant influence.
- Focus on assumptions with the highest impact.
17.2 Test Variability:
- Change key assumptions to see how the estimate varies.
- Document sensitivity ranges for these key variables.
17.3 Make Necessary Adjustments:
- Adjust the final estimate based on sensitivity findings.
- Ensure the final range accounts for these sensitivities.
</daily_estimation_process>
+++++
<daily_discussions>
Each day's discussion will have a specific focus:
- Day 1: Defining the core question
- Day 2: Listing all known facts among the estimators
- Day 3: Identifying key variables and considering potential biases or misconceptions
- Day 4: Decomposing the problem and making justifiable assumptions
- Day 5: Analogical transfer, with 3 academic experts attending for an extended discussion
- Day 6: Challenging assumptions
- Day 7: Sharing calculations from each day, reflecting on how estimations changed over the week, and modifying final answers if necessary
+++
Enrico Fermi will take extensive notes during every discussion and present a summary of all the estimators' work in the final <fermi_response> tags.
</daily_discussions>
+++++
>>>RESPONSE<<<
<output>
Present your solution for each estimator and each day using the following structure:

<day_n> (where n is the day number, 1-7)
  <estimator_m> (where m is the estimator number, 1-4)
    <thinking>
      [Plan your approach, considering the estimator's quartile bias and the day's focus]
    </thinking>
    <estimation_process>
      [Detailed steps of the estimation process]
    </estimation_process>
    <fermi_solution>
      [Final estimate and key points from the estimation process]
    </fermi_solution>
  </estimator_m>
  [Repeat for each estimator]
  <daily_discussion>
    [Summary of the day's discussion, focusing on the day's specific topic]
  </daily_discussion>
  <fermi_notes>
    [Enrico Fermi's notes and observations from the day's estimations and discussion]
  </fermi_notes>
</day_n>

After presenting all seven days of estimates and discussions, provide a final summary in <fermi_response> tags, comparing the different estimates, discussing the factors that led to variations between them, and how the estimates evolved over the week-long process.
</output>
+++++
<remember>
Remember to show your work clearly and explain your reasoning throughout each solution. Use logarithmic thinking and view quantities in terms of orders of magnitude. Incorporate meta-cognitive checks throughout your process to ensure you're avoiding biases and considering multiple perspectives.
</remember>
+++++
<week_of_estimation>
Begin your response by addressing the Fermi question for Day 1, starting with Estimator 1's estimation process. Iterate through each Estimator for each Day sequentially.
</week_of_estimation>
>>>END<<<

Summary and Takeaways

I hope you’ve enjoyed exploring these prompts! You may have noticed slight variations in syntax and style across the examples. This is intentional, as I test prompts across different LLMs to optimize for the best performance. For any aspiring prompt engineer, I recommend building a strong evaluation system and using observability tools to monitor model behaviour. Even a simple spreadsheet to log your prompt versions, model settings, and expected outcomes can be tremendously helpful in refining your approach.

These examples are just the beginning of what LLMs can do. As we continue to explore their potential, let’s remember to use these tools responsibly, keeping an eye on API costs and avoiding the temptation to “test” with unsuspecting family members!

Until next time,

Gareth

Leave a comment

Subscribe now

Share