Hallucination Detection on HaluEval Summarization (Starling-LM-7B-alpha)
81TPRKnowHalu
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| KnowHaluStructure=Structured, Time=8h query and 3h judge2025.12 | 81 | 43.4 | 62.2 | |
| PAPERStructure=Structured2025.12 | 80.2 | 45.4 | 62.8 | |
| KnowHaluStructure=Unstructured, Time=8h query and 3h judge2025.12 | 68.6 | 62 | 65.3 | |
| PAPERStructure=Unstructured2025.12 | 65 | 67.2 | 66.1 | |
| Knowhalu + HHEMStructure=Unstructured, Time=8h query and 10 min judge2025.12 | 54.4 | 69 | 61.7 | |
| Knowhalu + HHEMStructure=Structured, Time=8h query and 10 min judge2025.12 | 53 | 68.8 | 60.9 | |
| HHEMTime=10 min judge2025.12 | 32.2 | 79.4 | 55.8 |