Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Token-level Hallucination Detection on AIME 2025 (AUROC/AUPRC)

89.47AUROC

TOKENHD-8B

63.345270.127676.9183.6924May 12, 2026
Updated 21d ago

Evaluation Results

MethodLinks
2026.05
89.4780.14
2026.05
82.2660.88
2026.05
75.4349.87
2026.05
70.8252.12
64.3546.63