Hallucination Detection on TriviaQA French - High (test)
95.5AUROCMultiHaluDet
Evaluation Results
| Method | Links | |
|---|---|---|
| MultiHaluDetBackbone=Mistral-7B2026.05 | 95.5 | |
| MultiHaluDetBackbone=LLaMA2-7B2026.05 | 94.9 | |
| Best BaselineBackbone=Mistral-7B2026.05 | 81.3 | |
| Best BaselineBackbone=LLaMA2-7B2026.05 | 80.5 |