Hallucination Detection on E2E (test)
90F1-RPoolingE
Evaluation Results
| Method | Links | |
|---|---|---|
| PoolingEClassifier architecture=Ensembled, Training data distribution=Organic2023.12 | 90 | |
| PoolingSLClassifier architecture=Single-layer, Training data distribution=Organic2023.12 | 89 | |
| Expert Human2023.12 | 79 | |
| ChatGPTPrompting strategy=Standard2023.12 | 74 | |
| Seq-Logprob2023.12 | 72 | |
| FActScore2023.12 | 72 | |
| Optimized Coin (OC)Threshold optimization=Validation set tuned2023.12 | 71 | |
| PoolingE-SYNTHClassifier architecture=Ensembled, Training data distribution=Synthetic2023.12 | 71 | |
| PoolingSL-SYNTHClassifier architecture=Single-layer, Training data distribution=Synthetic2023.12 | 68 | |
| ChatGPT CoTPrompting strategy=Chain-of-thought2023.12 | 47 |