ResearchBenchmarksHallucination monitoring on Inheritance domainFollow90AUROCHalMit59.8467.6775.583.33Jul 21, 2025Evaluation ResultsMethodMethodLinksAUROCAUC-PRF1AccuracyHalMitBackbone=Llama3.1Backbone=Llama3.12025.0790868288SCGBackbone=Llama3.1Backbone=Llama3.12025.0785848585PPBackbone=Llama3.1Backbone=Llama3.12025.0771797271SCGBackbone=Llama2Backbone=Llama22025.0770736875HalMitBackbone=Llama2Backbone=Llama22025.0770798078ICLBackbone=Llama2Backbone=Llama22025.0769675956PPBackbone=Llama2Backbone=Llama22025.0768607775ICLBackbone=Llama3.1Backbone=Llama3.12025.0761736565