Hallucination Detection on TriviaQA Bangla Medium (test)
87.6AUROC (%)MultiHaluDet
Evaluation Results
| Method | Links | |
|---|---|---|
| MultiHaluDetBackbone=Mistral-7B2026.05 | 87.6 | |
| MultiHaluDetBackbone=LLaMA2-7B2026.05 | 86.3 | |
| Best BaselineBackbone=Mistral-7B2026.05 | 69.2 | |
| Best BaselineBackbone=LLaMA2-7B2026.05 | 67.8 |