List your company

Hallucination Detection on SelfAware GPT outputs (test)

0.528AUROCSonnet judge
0.323120.376310.42950.48269Jun 1, 2026

Evaluation results

MethodLinks
Sonnet judgeScorer Category=Judge2026.060.528
Gemini judgeScorer Category=Judge2026.060.513
Exact MatchScorer Category=BB2026.060.499
Semantic Neg.Scorer Category=BB2026.060.495
Sem. Sets Conf.Scorer Category=BB2026.060.493
Prob. MarginScorer Category=WB2026.060.483
Mean Token Neg.Scorer Category=WB2026.060.474
Cosine Sim.Scorer Category=BB2026.060.471
Sequence Prob.Scorer Category=WB2026.060.47
GPT-4o judgeScorer Category=Judge2026.060.467
Min Token Prob.Scorer Category=WB2026.060.44
Llama judgeScorer Category=Judge2026.060.409
Non-Cont.Scorer Category=BB2026.060.401
EntailmentScorer Category=BB2026.060.355
P(True)Scorer Category=WB2026.060.331