Hallucination Detection on CNN/DM
76.14AUROCFEPoID
Evaluation Results
| Method | Links | |
|---|---|---|
| FEPoIDModel=Mistral-7B-Instruct-v0.3, w=7, FST=false2026.05 | 76.14 | |
| CurvatureModel=Mistral-7B-Instruct-v0.3, w=7, FST=false2026.05 | 73.19 | |
| IDModel=Mistral-7B-Instruct-v0.3, w=7, FST=false2026.05 | 71.85 | |
| RGNModel=Mistral-7B-Instruct-v0.3, w=7, FST=false2026.05 | 70.31 | |
| Val LossModel=Mistral-7B-Instruct-v0.3, w=7, FST=false2026.05 | 69.38 | |
| RankMeModel=Mistral-7B-Instruct-v0.3, w=7, FST=false2026.05 | 68.69 | |
| SNRModel=Mistral-7B-Instruct-v0.3, w=7, FST=false2026.05 | 68.11 | |
| FEPoIDModel=LlaMA-3.1-8B-Instruct, w=7, FST=false2026.05 | 59.95 | |
| CurvatureModel=LlaMA-3.1-8B-Instruct, w=7, FST=false2026.05 | 59.22 | |
| IDModel=LlaMA-3.1-8B-Instruct, w=7, FST=false2026.05 | 59.18 | |
| Val LossModel=LlaMA-3.1-8B-Instruct, w=7, FST=false2026.05 | 58.59 | |
| RGNModel=LlaMA-3.1-8B-Instruct, w=7, FST=false2026.05 | 58.21 | |
| RankMeModel=LlaMA-3.1-8B-Instruct, w=7, FST=false2026.05 | 57.74 | |
| SNRModel=LlaMA-3.1-8B-Instruct, w=7, FST=false2026.05 | 54.74 |