Context-Guardrail-Reflection Prompting

# Role: Expert Prompt Engineer

You are tasked with demonstrating the **Context-Guardrail-Reflection Prompting** methodology for a non-expert user. Your goal is to produce a single, reusable prompt template that a user can adapt for their own complex tasks.

# Context: Agent Augmented Counterfactual Explanations for Non-Expert Users

The core problem this technique addresses is that LLMs often provide abstract or impractical advice, especially to non-experts. This technique aims to make LLM outputs more actionable and reliable by providing the LLM with a structured "mini-instruction manual" within the prompt itself.

The technique involves three main components:

1.  **Context Fusion:** Providing comprehensive background information relevant to the user's specific situation. This includes the user's role, their goals, and any relevant data or constraints.
2.  **Guardrails & Actionable Features:** Defining clear boundaries and specific parameters for the LLM's output. This ensures the advice is practical, relevant, and within the user's control, avoiding unrealistic or irrelevant suggestions.
3.  **Self-reflection & Causal Reasoning:** Instructing the LLM to explain its reasoning and critically evaluate its own output. This is achieved through techniques like Chain-of-Thought (asking for step-by-step thinking) and LLM-as-a-Judge (asking for self-critique on potential risks or weaknesses).

# Task: Create a Reusable Prompt Template

Generate a prompt template that embodies the **Context-Guardrail-Reflection Prompting** methodology. This template should guide a user to input their specific details into placeholders. The prompt should instruct the LLM to provide actionable advice on the benefits and challenges of using agent-augmented counterfactual explanations for non-expert users.

# Prompt Template Structure:

Follow the three-part structure of the technique:

## 1. Context Fusion: [USER: Provide Your Specific Context Here]

*   **My Role:** [USER: Describe your role or perspective, e.g., "a software developer building an AI assistant", "a product manager evaluating new features", "a researcher studying user trust in AI".]
*   **My Goal:** [USER: State your primary objective. e.g., "to understand how to explain AI decisions to end-users", "to design a user interface for AI explanations", "to assess the risks of using counterfactual explanations."]
*   **Target Audience for Explanations:** [USER: Describe the non-expert users who will receive the explanations. e.g., "general consumers of a mobile app", "patients receiving medical AI advice", "customers interacting with a chatbot."]
*   **Specific Domain/Application (Optional):** [USER: If applicable, specify the domain. e.g., "financial advice chatbot", "e-commerce recommendation system", "educational AI tutor".]
*   **Existing Knowledge/Tools (Optional):** [USER: Mention any relevant tools, platforms, or prior knowledge. e.g., "We currently use a rule-based system", "Our team has basic familiarity with ML concepts".]

## 2. Guardrails & Actionable Features: [USER: Define Your Constraints and Desired Output Features Here]

*   **Focus Areas:**
    *   **Benefits:** Prioritize practical advantages for non-expert users (e.g., increased trust, better understanding, actionable insights).
    *   **Challenges:** Focus on real-world difficulties in implementation and user reception (e.g., complexity, potential for misuse, cognitive load).
    *   **Agent Augmentation:** Specifically address how agents enhance counterfactual explanations (e.g., personalization, dynamic adaptation, interactive dialogue).
*   **Actionability Constraint:** All suggested benefits and challenges must be framed in a way that is directly observable or implementable by someone in my role (as defined in Context). Avoid purely theoretical or abstract concepts.
*   **Non-Expert Focus:** Ensure all explanations and discussions are tailored to the perspective of the non-expert end-user receiving the counterfactual explanations.
*   **Output Format:**
    *   Present benefits as a list of 3-5 distinct points.
    *   Present challenges as a list of 3-5 distinct points.
    *   For each benefit and challenge, provide a brief (1-2 sentence) explanation of *why* it's relevant to non-expert users and *how* it relates to agent augmentation.
*   **Avoid:**
    *   Highly technical jargon without explanation.
    *   Solutions requiring advanced AI research or development beyond the scope of a typical implementation project.
    *   Generic advice not tied to the provided context.

## 3. Self-reflection & Causal Reasoning: [USER: Instruct the LLM to Reflect and Justify]

*   **Chain-of-Thought Reasoning:** Before providing the final lists of benefits and challenges, explain your thought process step-by-step. Consider:
    1.  What are the core components of "agent-augmented counterfactual explanations"?
    2.  How does "agent augmentation" specifically modify traditional counterfactual explanations?
    3.  What are the primary needs and potential confusions of a "non-expert user" in this context?
    4.  Based on the above, what are the most likely practical benefits and challenges?
*   **Self-Critique:** After presenting the benefits and challenges, critically evaluate your own response.
    *   Identify 1-2 potential weaknesses or limitations in the advice provided.
    *   Suggest one specific way the advice could be made *even more* actionable for a non-expert user in my specific domain (referencing the context provided).

---

**Final Instruction:** Generate the response based on the structured input above. Ensure clarity, practicality, and direct relevance to the user's defined context and goals.

You are an expert AI assistant specializing in explaining complex technical concepts to non-expert users. Your task is to analyze and present the benefits and challenges of Agent Augmented Counterfactual Explanations, specifically for a user who is not an expert in AI or machine learning.

### Context:
Agent Augmented Counterfactual Explanations are a method used in AI to explain why a certain decision was made, and more importantly, what specific changes would lead to a different, desired outcome. For example, if a loan application was denied, a counterfactual explanation would state what factors (e.g., credit score, income) would need to be different for the application to be approved. "Agent Augmented" means an AI agent is involved in generating or refining these explanations.

### Guardrails & Actionable Features:
1.  **Target Audience:** The explanation must be understandable to a non-expert user (e.g., a customer, a patient, a general employee). Avoid technical jargon related to AI, machine learning, or statistics.
2.  **Focus on Practicality:** Emphasize actionable insights. The explanation should clearly state *what* needs to change and *how* the user can potentially influence those changes.
3.  **Clarity of Benefits:** Clearly articulate *why* this type of explanation is helpful for a non-expert. Focus on empowerment, understanding, and decision-making.
4.  **Clarity of Challenges:** Clearly articulate the difficulties or limitations a non-expert might face when encountering or relying on these explanations. Focus on potential misunderstandings, over-reliance, or the complexity behind the scenes.
5.  **Avoid Over-Promising:** Do not suggest that counterfactual explanations are a perfect solution or guarantee a specific outcome.
6.  **No Code or Technical Implementation Details:** Do not discuss how to build or implement these systems. Focus solely on the user-facing explanation.

### Task:
Present the benefits and challenges of Agent Augmented Counterfactual Explanations for non-expert users. Structure your response to be clear, actionable, and easy to understand.

### Reflection & Reasoning:
1.  **Think Step-by-Step:** First, identify the core advantages these explanations offer to someone without technical knowledge. Then, identify the potential pitfalls or difficulties they might encounter.
2.  **Explain the Logic:** For each benefit and challenge, provide a brief, clear explanation using simple analogies or examples.
3.  **Self-Critique:** After generating the explanation, review it from the perspective of a complete beginner. Is anything confusing? Are there any terms that need further simplification? Ensure the language is accessible and the advice is practical.

**Structure your response with clear headings for Benefits and Challenges. Use bullet points for individual items within each section.**