Hallucination Detection on HaluEval GPT outputs (test)
0.582AUROCP(True)
Evaluation Results
| Method | Links | |
|---|---|---|
| P(True)Scorer Category=WB2026.06 | 0.582 | |
| Non-Cont.Scorer Category=BB2026.06 | 0.579 | |
| Sem. Sets Conf.Scorer Category=BB2026.06 | 0.579 | |
| Mean Token Neg.Scorer Category=WB2026.06 | 0.579 | |
| GPT-4o judgeScorer Category=Judge2026.06 | 0.579 | |
| Prob. MarginScorer Category=WB2026.06 | 0.576 | |
| Semantic Neg.Scorer Category=BB2026.06 | 0.575 | |
| EntailmentScorer Category=BB2026.06 | 0.572 | |
| Cosine Sim.Scorer Category=BB2026.06 | 0.568 | |
| Sonnet judgeScorer Category=Judge2026.06 | 0.567 | |
| Sequence Prob.Scorer Category=WB2026.06 | 0.562 | |
| Gemini judgeScorer Category=Judge2026.06 | 0.562 | |
| Min Token Prob.Scorer Category=WB2026.06 | 0.528 | |
| Llama judgeScorer Category=Judge2026.06 | 0.518 | |
| Exact MatchScorer Category=BB2026.06 | 0.516 |