List your company

Hallucination Detection on SelfAware Gemini outputs (test)

52.8AUROCSonnet judge
36.26440.55744.8549.143Jun 1, 2026

Evaluation results

MethodLinks
Sonnet judgeScorer Category=Judge2026.0652.8
GPT-4o judgeScorer Category=Judge2026.0651.9
P(True)Scorer Category=WB2026.0651
Gemini judgeScorer Category=Judge2026.0650.2
Exact MatchScorer Category=BB2026.0649.9
Prob. MarginScorer Category=WB2026.0648.6
Mean Token Neg.Scorer Category=WB2026.0646.9
Sequence Prob.Scorer Category=WB2026.0646.9
Llama judgeScorer Category=Judge2026.0644.5
Cosine Sim.Scorer Category=BB2026.0644.4
Non-Cont.Scorer Category=BB2026.0642.7
Semantic Neg.Scorer Category=BB2026.0639.3
Min Token Prob.Scorer Category=WB2026.0639
Sem. Sets Conf.Scorer Category=BB2026.0638.7
EntailmentScorer Category=BB2026.0636.9