Edit Step4_qc_link Template

Step4_qc_link v2

Viewer

Test Case:

Template

Preview

History

System prompt (optional) User prompt (template)

You are analyzing NSPE Board of Ethical Review questions and conclusions to determine which conclusion answers which question.

**QUESTIONS:**
{{ questions_text }}

**CONCLUSIONS:**
{{ conclusions_text }}

**TASK:**
For each conclusion, determine which question(s) it answers.

Consider:
- Direct answers: Conclusion explicitly addresses the question
- Partial answers: Conclusion addresses part of a multi-part question
- Implicit answers: Conclusion answers question without restating it

**OUTPUT FORMAT (JSON):**
```json
[
  {
    "conclusion_number": 1,
    "answers_questions": [1],
    "confidence": 0.95,
    "reasoning": "Conclusion 1 directly addresses Question 1 by stating whether Engineer A violated II.4.e, which is exactly what Question 1 asked."
  },
  {
    "conclusion_number": 2,
    "answers_questions": [1, 2],
    "confidence": 0.90,
    "reasoning": "Conclusion 2 addresses both Question 1 (violation determination) and Question 2 (disclosure requirement) as related issues."
  }
]
```

**IMPORTANT:**
- A conclusion can answer multiple questions
- Multiple conclusions can answer the same question
- Provide confidence (0.0-1.0) and reasoning for each link
- If a conclusion doesn't answer any question, use answers_questions: []

Loading preview...

Version History

No version history yet. Save changes to create versions.

Saved

1255 chars

Reference

Template Variables

Numbered questions, each as a **Question N:** header plus quoted text

Numbered conclusions, each as a **Conclusion N:** header plus quoted text and a Type line

Output Schema

format json_array

item_fields

JSON Wrapper

Appended to the prompt by the extractor:

IMPORTANT: Wrap your response as a JSON object with exactly two keys:
{"new_step4_qc_link_classes": [...], "step4_qc_link_individuals": [...]}
If there are no individuals to report, use an empty array for "step4_qc_link_individuals".