## Screenshot: Instance Reflection Query Template ### Overview The image displays a structured template for comparing a correct solution and a potential solution to a math problem. It includes explicit instructions for error identification, formatting requirements, and verification steps. ### Components/Axes - **Title**: "Instance reflection query template" (top-left, bold black text). - **User Section**: Contains step-by-step instructions for comparing solutions. - **Placeholders**: - `{{ Question }}` - `{{ correct_solution }}` - `{{ potential_solution }}` - **Verification Prompt**: "Verification: Is the previous solution correct? (Yes/No)" - **Explanation Guideline**: Requires a specific phrase ("I think the solution is correct.") if the potential solution is valid. ### Detailed Analysis 1. **User Instructions**: - Compare two solutions to the same math question. - Identify errors in the potential solution (if any). - Explain the error in **two sentences or less** within `<reflection>` tags. - Conclude with a verification question. - Prohibits solving the question or implying comparisons between solutions. 2. **Formatting Rules**: - Use `<reflection>` tags for error explanations. - Avoid phrases like "Compare..." or "Difference between...". 3. **Placeholders**: - `{{ Question }}`: Replaced with the actual math problem. - `{{ correct_solution }}`: The accurate solution. - `{{ potential_solution }}`: The solution to be evaluated. ### Key Observations - The template enforces strict formatting (e.g., `<reflection>` tags, no comparative language). - The verification step is binary (Yes/No), with a predefined explanation for correct solutions. - Placeholders are enclosed in double curly braces (`{{ }}`), indicating variables to be filled. ### Interpretation This template is designed for automated or semi-automated evaluation of solutions, likely in an educational or AI-driven context. The emphasis on brevity (`<reflection>` tags) and prohibition of comparative language suggests it prioritizes concise, direct error analysis. The verification step acts as a final check, ensuring consistency in grading or feedback. The structure implies integration with a system that parses these placeholders and instructions programmatically. No numerical data, charts, or diagrams are present. The focus is entirely on textual instructions and placeholders.