List your company

Semantic Hallucination Detection on PAWS

91.24AUROCHALLUGUARD
63.513670.711877.9185.1082Jan 26, 2026

Evaluation results

MethodLinks
HALLUGUARDBase Model=QwQ-32B2026.0191.2488.53
HALLUGUARDBase Model=Llama2-70B2026.0190.1887.64
RACEBase Model=QwQ-32B2026.0186.3883.41
HALLUGUARDBase Model=Llama-3.2-3B2026.0185.6382.14
InsideBase Model=Llama2-70B2026.0185.4782.38
InsideBase Model=QwQ-32B2026.0185.4182.27
MINDBase Model=QwQ-32B2026.0184.5681.37
RACEBase Model=Llama2-70B2026.0184.2681.73
MINDBase Model=Llama2-70B2026.0183.9281.06
SelfCheckGPTBase Model=QwQ-32B2026.0183.5979.42
SelfCheckGPTBase Model=Llama2-70B2026.0182.7178.89
InsideBase Model=Llama-3.2-3B2026.0180.4677.28
Semantic EntropyBase Model=QwQ-32B2026.0180.4277.18
RACEBase Model=Llama-3.2-3B2026.0179.4776.28
Semantic EntropyBase Model=Llama2-70B2026.0179.0676.32
MINDBase Model=Llama-3.2-3B2026.0178.9375.41
P(true)Base Model=QwQ-32B2026.0178.5375.21
EnergyBase Model=QwQ-32B2026.0178.2975.44
SelfCheckGPTBase Model=Llama-3.2-3B2026.0177.8273.41
LN EntropyBase Model=QwQ-32B2026.0177.4373.29
P(true)Base Model=Llama2-70B2026.0177.3974.18
EnergyBase Model=Llama2-70B2026.0177.1474.28
PerplexityBase Model=QwQ-32B2026.0176.7272.63
LN EntropyBase Model=Llama2-70B2026.0176.2372.59
PerplexityBase Model=Llama2-70B2026.0175.6871.42
Semantic EntropyBase Model=Llama-3.2-3B2026.0175.1172.41
P(true)Base Model=Llama-3.2-3B2026.0173.5670.43
FActScoreBase Model=QwQ-32B2026.0173.4668.32
EnergyBase Model=Llama-3.2-3B2026.0173.0570.22
FActScoreBase Model=Llama2-70B2026.0172.6267.58
LN EntropyBase Model=Llama-3.2-3B2026.0172.1968.34
PerplexityBase Model=Llama-3.2-3B2026.0171.2767.55
Lexical SimilarityBase Model=QwQ-32B2026.0169.5464.27
FActScoreBase Model=Llama-3.2-3B2026.0168.4463.58
Lexical SimilarityBase Model=Llama2-70B2026.0168.3563.44
Lexical SimilarityBase Model=Llama-3.2-3B2026.0164.5859.67