PUBLIC EMPLOYER DEMONSTRATION · V2.7
Evidence Review & Grading Laboratory
Retrieve one of 20 synthetic prompt examples, inspect both illustrative model responses, independently score them, and export a reviewer report. This is a local, browser-only scoring exercise: no private data, model API calls, or server-side submissions.
◈ Prompt evidence
Original prompt
Expected outcome
Claude · sample response
Illustrative Authenticator findings
Suggested improvement
⚖️ Independent reviewer scorecard
Interactive · browser-onlyGrade each response from 0 to 5 for every criterion. Weighted totals update immediately. Record your critique and export the review as JSON. Scores are reset when you select another record.
ChatGPT review
Claude review
🛡 Security & evaluation integrity
No authenticated archive or private files are requested.
All included examples are labeled synthetic.
Independent reviewer grades are separate from example scores.
Report is generated locally; nothing is submitted to the server.
Production evaluation requires source provenance, reproducible test suites, real provider/version settings, authenticated reviews, and server-side audit records. This page does not claim to implement those production safeguards.
