Hallucination Detection on HaluEval Bangla - Medium (test)
89.1AUROCMultiHaluDet
Evaluation Results
| Method | Links | |
|---|---|---|
| MultiHaluDetBackbone=Mistral-7B2026.05 | 89.1 | |
| MultiHaluDetBackbone=LLaMA2-7B2026.05 | 88.4 | |
| Best BaselineBackbone=Mistral-7B2026.05 | 78.4 | |
| Best BaselineBackbone=LLaMA2-7B2026.05 | 75.1 |