Hallucination Detection on HalluQA
94.44AccuracyANAH-v2-Stage3
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ANAH-v2-Stage3Method=Zero-Shot2024.07 | 94.44 | — | |
| ANAH-v2-Stage2Method=Zero-Shot2024.07 | 92.63 | — | |
| ANAH-v2-Stage1Method=Zero-Shot2024.07 | 91.74 | — | |
| GPT4Method=Zero-Shot2024.07 | 62.81 | — | |
| HalluCleanBackbone=GPT-3.5-turbo2025.11 | 55 | 41.6 | |
| GPT-3.5-turboBackbone=GPT-3.5-turbo2025.11 | 46.5 | 7 | |
| GAME-LoRAProtocol=Training-time2026.01 | 44.5 | — | |
| CADProtocol=Inference-time2026.01 | 41.7 | — | |
| MEProtocol=Training-time2026.01 | 39.5 | — | |
| DisagreementProtocol=Training-time2026.01 | 39.1 | — | |
| BaselineArchitecture=LoRA, Loss=Vanilla CE, Protocol=Training-time2026.01 | 37.6 | — | |
| ActDecProtocol=Inference-time2026.01 | 37.6 | — |