model data_extraction template risk: low
Document Claim Verifier with Quotes
Instructs the model to base answers only on uploaded documents, cite exact quotes for each claim using [Document, Page/Section, Quote] format, state 'Not found' or [Unverified] if…
PROMPT
1. Base your answer ONLY on the uploaded documents. Nothing else. 2. If info isn't found, say "Not found." Don't guess. 3. For each claim, cite: [Document, Page/Section, Quote] 4. If uncertain, mark as [Unverified] 5. [Your question] Re-scan the document. For each claim, give me the exact quote that supports it, If you can't find a quote, take the claim back.
INPUTS
- question REQUIRED
-
Placeholder for the specific question or claims to verify
e.g. [Your question]
REQUIRED CONTEXT
- uploaded documents
- question/claims
ROLES & RULES
- Base your answer ONLY on the uploaded documents. Nothing else.
- If info isn't found, say "Not found." Don't guess.
- For each claim, cite: [Document, Page/Section, Quote]
- If uncertain, mark as [Unverified]
- Re-scan the document.
- For each claim, give me the exact quote that supports it.
- If you can't find a quote, take the claim back.
EXPECTED OUTPUT
- Format
- structured_report
- Constraints
-
- Base answer ONLY on uploaded documents
- If info not found say 'Not found'
- For each claim cite [Document, Page/Section, Quote]
- If uncertain mark [Unverified]
- Provide exact quote or retract claim
SUCCESS CRITERIA
- Verify claims using only uploaded documents
- Provide exact quotes for each claim
- Cite sources in [Document, Page/Section, Quote] format
- Say "Not found" or retract unsupported claims
FAILURE MODES
- Using information outside uploaded documents
- Guessing or fabricating details
- Failing to cite sources properly
- Not marking uncertain claims as [Unverified]
CAVEATS
- Dependencies
-
- Uploaded documents
- Specific question or claims to verify
- Missing context
-
- Explicit output structure for handling multiple claims.
- Examples of expected citation format.
- Method for identifying claims within the question.
- Ambiguities
-
- "Claims" not explicitly defined; assumed to be in the question.
- "take the claim back" is ambiguous (retract? deny?).
- Inconsistency between plural "documents" and singular "document".
- Repetitive instructions between numbered list and final paragraph.
QUALITY
- OVERALL
- 0.80
- CLARITY
- 0.80
- SPECIFICITY
- 0.85
- REUSABILITY
- 0.90
- COMPLETENESS
- 0.75
IMPROVEMENT SUGGESTIONS
- Merge repetitive parts into a single, numbered list of instructions.
- Specify a structured output format, e.g., 'Claim 1: [text]\nSupport: [quote or "Not found"]\nCitation: [Document, Page/Section]'
- Clarify "take the claim back" to "state 'Not supported' and explain why".
- Use consistent terminology: always "documents" plural.
- Replace "[Your question]" with a template placeholder like "{question}".
USAGE
Copy the prompt above and paste it into your AI of choice — Claude, ChatGPT, Gemini, or anywhere else you're working. Replace any placeholder sections with your own context, then ask for the output.
MORE FOR MODEL
- Job Posting Snapshot Preservation Enginemodeldata_extraction
- Natural Language SQL Query Generatormodeldata_extraction
- LinkedIn JSON to Canonical Markdown Profile Generatormodeldata_extraction
- Visual Clutter Text Cleanermodeldata_extraction
- Company Shareholder JSON Extractormodeldata_extraction
- PDF to GitHub Markdown Convertermodeldata_extraction
- Vision-to-JSON Image Analyzermodeldata_extraction
- Chat Transcript Exporter with Reversalmodeldata_extraction
- Model Parameters Table Image to CSVmodeldata_extraction
- Webpage Parser with Embed Handling and Translationmodeldata_extraction