List your company

Hallucination Detection on HaluEval held-out 50% (test)

69.9AUROCLlama-3-8B
51.756.42561.1565.875Jun 1, 2026

Evaluation results

MethodLinks
Llama-3-8BEnsemble Strategy=Best J + Best BB2026.0669.9
Llama-3-8BEnsemble Strategy=Best J2026.0668.7
Llama-3-8BEnsemble Strategy=Weighted Avg.2026.0668.3
Llama-3-8BEnsemble Strategy=Best J + Best WB2026.0666.1
GPT-4oEnsemble Strategy=Best J + Best BB2026.0658.1
Gem.-1.5-FlEnsemble Strategy=Best J + Best BB2026.0658.1
GPT-4oEnsemble Strategy=Best J2026.0657
Gem.-1.5-FlEnsemble Strategy=Weighted Avg.2026.0656.8
GPT-4oEnsemble Strategy=Weighted Avg.2026.0656.2
Gem.-1.5-FlEnsemble Strategy=Best J2026.0654.9
GPT-4oEnsemble Strategy=Best J + Best WB2026.0654.2
Gem.-1.5-FlEnsemble Strategy=Best J + Best WB2026.0652.4