List your company

Faithfulness Hallucination Detection on LLM-AggreFact Refined

86.8Agg-CNNDeepSeek-V3.2
44.05655.15366.2577.347Dec 23, 2025

Evaluation results

MethodLinks
DeepSeek-V3.2thinking=true2025.1286.876.88880.78877.485.992.183.591.481.8
FaithLensbackbone=Llama-3.1-8B-Inst2025.1284.97989.479.692.492.186.892.285.187.285.6
FactCGparameters=435M2025.1276.968.176.275.38986.579.69079.171.972.2
Claude-3.7-Sonnet2025.1275.673.683.774.486.986878885.48486
DeepSeek-V3.2thinking=false2025.1275.565.575.474.487.972.980.89165.582.972.7
GPT-4.12025.1274.173.681.680.391.381.189.193.275.986.386.4
ClearCheckbackbone=Llama-3.1-8B-Inst2025.1272.878.685.472.787.98783.88767.875.881.8
MiniCheckparameters=7B2025.127072.785.672.986.88986.99174.377.885.9
o12025.1268.376.777.172.38576.179.685.965.876.178.7
o32025.1267.877.283.379.686.987.780.692.282.983.882.1
Llama-3.1-405B-Inst2025.1265.571.680.768.88276.380.78467.378.872.5
o3-mini2025.1264.481.580.2738681.580.784.877.178.384
GPT-4o2025.1262.374.978.368.3867581.886.971.576.977.9
AlignScoreparameters=355M2025.1245.76879.87583.786.683.692.275.876.567.3