List your company

Hallucination Detection on HaluEval Llama outputs (test)

0.704AUROCGPT-4o judge
0.48040.538450.59650.65455Jun 1, 2026

Evaluation results

MethodLinks
GPT-4o judgeScorer Category=Judge2026.060.704
EntailmentScorer Category=BB2026.060.656
Gemini judgeScorer Category=Judge2026.060.649
Non-Cont.Scorer Category=BB2026.060.64
P(True)Scorer Category=WB2026.060.638
Sem. Sets Conf.Scorer Category=BB2026.060.637
Semantic Neg.Scorer Category=BB2026.060.636
Cosine Sim.Scorer Category=BB2026.060.605
Sonnet judgeScorer Category=Judge2026.060.599
Sequence Prob.Scorer Category=WB2026.060.574
Llama judgeScorer Category=Judge2026.060.567
Mean Token Neg.Scorer Category=WB2026.060.559
Min Token Prob.Scorer Category=WB2026.060.557
Prob. MarginScorer Category=WB2026.060.535
Exact MatchScorer Category=BB2026.060.489