Hallucination Detection on HaluEval Llama outputs (test)
0.704AUROCGPT-4o judge
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-4o judgeScorer Category=Judge2026.06 | 0.704 | |
| EntailmentScorer Category=BB2026.06 | 0.656 | |
| Gemini judgeScorer Category=Judge2026.06 | 0.649 | |
| Non-Cont.Scorer Category=BB2026.06 | 0.64 | |
| P(True)Scorer Category=WB2026.06 | 0.638 | |
| Sem. Sets Conf.Scorer Category=BB2026.06 | 0.637 | |
| Semantic Neg.Scorer Category=BB2026.06 | 0.636 | |
| Cosine Sim.Scorer Category=BB2026.06 | 0.605 | |
| Sonnet judgeScorer Category=Judge2026.06 | 0.599 | |
| Sequence Prob.Scorer Category=WB2026.06 | 0.574 | |
| Llama judgeScorer Category=Judge2026.06 | 0.567 | |
| Mean Token Neg.Scorer Category=WB2026.06 | 0.559 | |
| Min Token Prob.Scorer Category=WB2026.06 | 0.557 | |
| Prob. MarginScorer Category=WB2026.06 | 0.535 | |
| Exact MatchScorer Category=BB2026.06 | 0.489 |