Our new X account is live! Follow @wizwand_team for updates

Hallucination Detection on HaluBench (test)

86.96HE

Lynx-8B

Updated 4d ago

Evaluation Results

Method	Links
Lynx-8B 2025.12		86.96	60.25	74.3	64.79	97.7	88.37	68.39	77.25
HDM-4B-RL 2025.12		83.96	68.52	84.54	74.39	92.5	77.79	75.88	79.65
o4-mini 2025.12		83.23	71.89	88.9	76.28	94.19	85.95	77.88	82.62
GPT-4.1 2025.12		82.95	62.58	60.79	66.91	90.75	88.17	75.41	75.37
HDM-4B-RL 2025.12		82.84	63.74	57.83	50.85	84.21	80.67	70.75	70.13
Qwen3-32B 2025.12		80.92	57.83	69.59	70.61	83.94	83.43	72.61	74.13
Bespoke-MiniCheck-7B 2025.12		79.73	53.08	50.38	64.78	88.1	64.75	78.56	68.48
Qwen3-32B 2025.12		77.63	69.36	86.58	80.1	92.99	87.1	78.16	81.7
Qwen3-4B 2025.12		75.15	61.28	82.98	76.19	87.61	77.72	76.26	76.74
Alignscore-Large 2025.12		74.1	56.77	41.93	50.9	64.11	61.05	65.6	59.21
Qwen3-4B 2025.12		73.64	58.25	56.72	52.29	78.26	68.83	66.43	64.92
MiniCheck-Flan-T5-Large 2025.12		66.85	53.65	42.46	49.74	75.48	71.72	76.2	62.3
HHEM 2025.12		66.59	73.72	40.85	46.55	57.48	57.22	61	57.63
FactCG-DeBERTa-v3-Large 2025.12		66.29	52.28	42.03	48.67	78.48	59.56	61	58.33