Compliance
AI safety and robustness evaluation summary — RITA
Results of RITA's latest evaluation campaign: prompt injection, leakage, accuracy, fairness, and continuous oversight.
REVIEWIT_EVALUATION-ROBUSTESSE-IA_Public_v1.0_FR.pdfFrench · authoritative · 103 KB · PDFDownloadREVIEWIT_EVALUATION-ROBUSTESSE-IA_Public_v1.0_EN.pdfEnglish · reviewed translation · 103 KB · PDFDownload
SHA-256 hash published in the registry: verifiable by you, without us. Verify →
Published version history
- 1.03 Nov 2026CURRENT