Response-level Hallucination Detection on RAGognize
94.91AUROCRAGognizer
Evaluation Results
| Method | Links | |
|---|---|---|
| RAGognizerDetector category=White-box, Proxy generator=Qwen3-4B-Instruct-2507, Same-dataset evaluation=true2026.04 | 94.91 | |
| LookbackLensDetector category=White-box, Proxy generator=Qwen3-4B-Instruct-2507, Same-dataset evaluation=true2026.04 | 94.36 | |
| MiniCheck-7BDetector category=Black-box2026.04 | 83.48 | |
| HallucinationProbesDetector category=White-box, Probe=LoRA-KL probe on LLaMA-3.1-8B (lambda_KL = 0.05)2026.04 | 83.25 | |
| HDM-2-3BDetector category=Black-box2026.04 | 75.41 | |
| SelfCheckGPT (NLI)Detector category=Black-box2026.04 | 74.22 | |
| HaloScopeDetector category=White-box, Proxy generator=Qwen3-4B-Instruct-2507, Same-dataset evaluation=true2026.04 | 71.38 | |
| DeBERTa-v3 (Ent.)Detector category=Black-box2026.04 | 68.21 | |
| LettuceDetect-LDetector category=Black-box2026.04 | 65.75 | |
| INSIDE (EigenScore)Detector category=White-box, Proxy generator=Qwen3-4B-Instruct-25072026.04 | 64.49 | |
| DeBERTa-v3 (Con.)Detector category=Black-box2026.04 | 64.31 | |
| HHEM-2.1-OpenDetector category=Black-box2026.04 | 61.49 | |
| Semantic EntropyDetector category=White-box2026.04 | 61.48 |