The Gray Zone of Faithfulness: Taming Ambiguity in Unfaithfulness Detection
信仰的灰色地带:在不忠检测中消除歧义
机构 * Key Lab of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences (CAS)(中国科学院信息处理重点实验室) ; State Key Lab of Al Safety(人工智能安全国家重点实验室) ; University of Chinese Academy of Sciences(中国科学院大学)
专题命中 评测与基准 :LLM(abstract);large language model(abstract);language model(abstract);分类 cs.CL、cs.AI
AI总结 本文提出了一种新的忠实性注释框架和VeriGray基准,用于解决摘要任务中因外部知识边界不明确导致的注释歧义问题,并展示了其对现有方法的挑战。
Comments Update the evaluation results due to the annotation updates; revise the citation to Seo et al., 2025; add the acknowledgements