R2VC：检索、验证与置信度校准的模块化事实核查
原标题：R2VC: Modular Fact-Checking with Retrieval, Verification, and Confidence Calibration
AI 摘要
研究者提出 R2VC，一种模块化的事实核查架构，将检索、推理、验证与置信度校准分离，结合稀疏+稠密混合检索、经 SFT 与 DPO 对齐的生成器、外部 NLI 交叉编码器筛选候选以及轻量级序列级校准器。在 FEVER 上，8B 骨干模型使用 R2VC 后准确率比基线提升 13.74%。消融实验显示验证器候选筛选与置信度校准贡献最大，移除候选筛选准确率降至 76.24%，移除校准使 Brier 分数近乎翻倍至 0.161；对 250 个错误的人工分析表明检索失败（尤其是错误实体证据）仍是主要瓶颈。
正文节选
R2VC: Modular Fact-Checking with Retrieval, Verification, and Confidence Calibration Abstract Large language models are increasingly used for automated fact checking, but end-to-end prompting often entangles evidence retrieval, reasoning, and uncertainty estimation, making failures difficult to diagnose and confidence difficult to trust. We present R2VC, a modular retrieve, reason, verify, calibrate architecture for evidence-grounded fact checking with citations and abstention. R2VC combines hyb