构建金融问答幻觉检测基准，评估知识图谱增强系统可靠性。
FinReflectKG -- HalluBench: GraphRAG Hallucination Benchmark for Financial Question Answering Systems
- 基于美国证券交易委员会10-K文件构建带标注的幻觉检测数据集。
- 嵌入方法在噪声三元组下仍保持9%性能下降，优于其他方法。
- 为金融、医疗等高风险领域提供AI可信性评估框架。
随着组织将AI问答系统用于合规、风险评估和决策支持，确保生成内容的事实准确性成为关键工程挑战。现有知识图谱（KG）增强型问答系统缺乏系统性幻觉检测机制——即事实错误输出，会损害可靠性并削弱用户信任。本文提出FinBench-QA-Hallucination，一个针对证券10-K文件的金融问答幻觉检测基准。数据集包含755个标注样本，来自300页文档，每条均通过严格证据链协议标注，要求文本段落与提取的关系三元组同时支持。评估六种检测方法：大模型裁判、微调分类器、自然语言推理模型、跨度检测器及基于嵌入的方法，在有无知识图谱三元组两种条件下进行测试。结果显示，大模型裁判与嵌入方法在干净条件下表现最佳（F1: 0.82–0.86）；但引入噪声三元组后，多数方法性能显著下降，马修斯相关系数（MCC）下降44%-84%，而嵌入方法仅下降9%。统计检验（Cochran's Q 和 McNemar）证实差异显著（p < 0.001）。研究揭示当前KG增强系统的脆弱性，为构建可靠金融信息系统提供依据，因幻觉可能导致监管违规与错误决策。该基准也为医疗、法律、政府等高风险领域提供可复用的AI可靠性评估框架。
原文摘要 · Abstract (English)
As organizations increasingly integrate AI-powered question-answering systems into financial information systems for compliance, risk assessment, and decision support, ensuring the factual accuracy of AI-generated outputs becomes a critical engineering challenge. Current Knowledge Graph (KG)-augmented QA systems lack systematic mechanisms to detect hallucinations - factually incorrect outputs that undermine reliability and user trust. We introduce FinBench-QA-Hallucination, a benchmark for evaluating hallucination detection methods in KG-augmented financial QA over SEC 10-K filings. The dataset contains 755 annotated examples from 300 pages, each labeled for groundedness using a conservative evidence-linkage protocol requiring support from both textual chunks and extracted relational triplets. We evaluate six detection approaches - LLM judges, fine-tuned classifiers, Natural Language Inference (NLI) models, span detectors, and embedding-based methods under two conditions: with and without KG triplets. Results show that LLM-based judges and embedding approaches achieve the highest performance (F1: 0.82-0.86) under clean conditions. However, most methods degrade significantly when noisy triplets are introduced, with Matthews Correlation Coefficient (MCC) dropping 44-84 percent, while embedding methods remain relatively robust with only 9 percent degradation. Statistical tests (Cochran's Q and McNemar) confirm significant performance differences (p < 0.001). Our findings highlight vulnerabilities in current KG-augmented systems and provide insights for building reliable financial information systems, where hallucinations can lead to regulatory violations and flawed decisions. The benchmark also offers a framework for integrating AI reliability evaluation into information system design across other high-stakes domains such as healthcare, legal, and government.
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。