基于对抗反馈的实时事实性评估
Real-time Factuality Assessment from Adversarial Feedback
AI总结：
提出利用基于 RAG 检测器的自然语言反馈，将实时新闻迭代改写为对抗性欺骗变体，以评估 LLM 检索证据并推理时事的能力。
摘要：
我们表明，现有用于评估传统来源新闻事实性的评测（例如事实核查网站上的声明），会使基于 LLM 的检测器随时间推移取得很高准确率，即使在其知识截止日期之后也是如此。这表明，来自此类来源的近期流行虚假信息之所以能被轻易识别，可能是因为它们出现在预训练/检索语料库中，或是这些数据集中出现了显著但浅层的模式。相反，我们认为，恰当的事实性评估数据集应当测试模型通过检索和阅读相关证据来对时事进行推理的能力。为此，我们开发了一种新型流水线，利用来自基于 RAG 的检测器的自然语言反馈，将实时新闻迭代改写为可挑战 LLM 的欺骗性变体。对于一个强大的、基于 RAG 的 GPT-4o 检测器，我们的迭代重写使二分类 ROC-AUC 绝对下降了 17.5 个百分点。实验揭示了 RAG 在评估和生成具有挑战性的新闻样本方面的重要作用：无检索的 LLM 检测器容易受到未见事件和对抗攻击影响，而来自基于 RAG 评估的反馈有助于发现更多欺骗性模式。
英文摘要：
We show that existing evaluations for assessing the factuality of news from conventional sources, such as claims on fact-checking websites, result in high accuracies over time for LLM-based detectors-even after their knowledge cutoffs. This suggests that recent popular false information from such sources can be easily identified due to its likely presence in pre-training/retrieval corpora or the emergence of salient, yet shallow, patterns in these datasets. Instead, we argue that a proper factuality evaluation dataset should test a model's ability to reason about current events by retrieving and reading related evidence. To this end, we develop a novel pipeline that leverages natural language feedback from a RAG-based detector to iteratively modify real-time news into deceptive variants that challenge LLMs. Our iterative rewrite decreases the binary classification ROC-AUC by an absolute 17.5 percent for a strong RAG-based GPT-4o detector. Our experiments reveal the important role of RAG in both evaluating and generating challenging news examples, as retrieval-free LLM detectors are vulnerable to unseen events and adversarial attacks, while feedback from RAG-based evaluation helps discover more deceitful patterns.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
