让AI在证据不足时选择不回答，提升事实核查可靠性。
Calibrated Selective Fact-Checking via Evidence Chain Evaluation
- 通过证据链评估机制，允许AI在不确定时选择跳过判断。
- 在95个测试案例中，正确回答93.7%的题目，6个弱证据案例被合理跳过。
- 适合需要高可信度输出的场景，如医疗、法律等关键领域。
大型语言模型虽能实现较强的事实核查准确率，但强制二元判断隐藏了可靠性问题：即使证据薄弱、稀疏或内部矛盾，系统仍可能给出自信结论。为此，我们提出证据链评估（ECE）框架，支持在不确定时选择不回答，而非对每个命题强行给出真/假判断。该系统是一个使用工具的验证代理，通过网络搜索、学术检索和可执行检查收集证据，并返回带有置信度与来源级元数据的结构化结论。在ECE-Bench上，ECE实现91.6%的标准准确率、93.7%的覆盖范围，以及在已回答命题上97.8%的选择性准确率。尽管在整体校准指标（如期望校准误差、Brier得分、AURC）上未超越最强检索基线，但其展现出明确的择优预测权衡：系统在已回答命题上保持极高准确率，同时有6/95个案例被合理跳过，这些案例集中于低可靠性证据层级（6例中有5例为源级别L4），表明弃权可作为处理认知薄弱证据的安全机制。代码已开源。
原文摘要 · Abstract (English)
Large language models (LLMs) can achieve strong fact-checking accuracy, yet forced binary decisions conceal a critical reliability problem: systems may issue confident verdicts even when supporting evidence is weak, sparse, or internally inconsistent. We address this issue through Evidence Chain Evaluation (ECE), a selective fact-checking framework that permits abstention via an uncertain verdict instead of requiring a true/false decision for every claim. The evaluated system is a tool-using verification agent that gathers evidence through web search, scholarly search, and executable checks, and then returns a structured verdict with confidence and source-level metadata. On ECE-Bench, ECE achieves 91.6% standard accuracy, 93.7% coverage, and 97.8% selective accuracy on answered claims. Although ECE does not outperform the strongest retrieval baseline on aggregate calibration metrics such as Expected Calibration Error, Brier score, or AURC, it delivers a clear selective-prediction trade-off: the system maintains very high accuracy on answered claims while deferring 6 of 95 cases. These deferred cases are concentrated in lower-reliability evidence settings (5/6 at source level L4), supporting the view that abstention functions as a safety-oriented mechanism for handling epistemically weak evidence. Code is available at https://github.com/ cheshireyang/ECE.git
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。