Spaces:
Running
Running
init: AX Diagnostic Leaderboard MVP — Chitos-pattern shell, 4-step menu (Visualize/Diagnose/Report/Remediate), 7 axes, seeded leaderboard (metacog trap_rate + serving live values)
4a0c678 | title: AX Diagnostic Leaderboard | |
| emoji: 🩺 | |
| colorFrom: blue | |
| colorTo: purple | |
| sdk: docker | |
| pinned: true | |
| license: apache-2.0 | |
| short_description: LLM AX diagnostic leaderboard across 7 axes | |
| # 🩺 AX Diagnostic Leaderboard | |
| VIDRAFT's AI Transformation (AX) diagnostic framework. Diagnoses LLMs like a clinical exam and publishes the scores. | |
| **Menu flow:** Visualize → Diagnose → Report → Remediate | |
| **7 evaluation axes:** Reliability/Metacognition · Intelligence · Structural Health (XRAY) · Serving Speed · Regulatory · Robustness · Fairness/Privacy → composite **AX Index (0–100, grade S/A/B/C/D)**. | |
| Aligned with international standards (Stanford HELM · TrustLLM · MLPerf Inference · NIST AI RMF · Korea AI Framework Act). Results and grades are public; diagnostic methods (patent-pending Model-XRAY, metacognition adapters) are proprietary. | |
| Operated by FINAL-Bench as a neutral benchmark operator. | |