arXiv:2608.22483v1 Announce Type: new
Abstract: Large Language Models (LLMs) increasingly support decision-making in high-stakes domains, but they often hallucinate and express confidence that is mis...
By Toghrul Abbasli, Kentaroh Toyoda, Yuan Wang, Li Chen
The paper introduces Fact-Ablated Evaluation (FAE), a framework that iteratively removes cited evidence to test whether large language models (LLMs) adjust their fact‑checking predictions accordingly. Experiments reveal that many off‑the‑shelf LLMs rely more on internal knowledge than on the provided evidence. To address this, the authors propose REAL, a training method that uses counterfactual evidence supervision to encourage LLMs to base veracity judgments on evidence, achieving better evidence‑dependent performance across four datasets.
By Xingyu Deng, Mingzi Cao, Nikolaos Aletras, Xi Wang, Mark Stevenson
ElementCheck is a new framework for evaluating the factuality of long-form text that addresses limitations of the traditional decompose‑retrieve‑verify pipeline. Rather than breaking sentences into atomic sub‑claims, it extracts entity pairs linked by verifiable connections to form an element graph, using the graph’s topology to gauge sentence complexity. This allows simple sentences to be verified directly while complex ones undergo targeted element‑level refinement, and the authors introduce the FastFact‑Sent benchmark to support fine‑grained evaluation, demonstrating consistent improvements across multiple backbone models.
By Xinming Wang, Haoran Du, Yi Chen, Jian Xu, Hongming Yang, Han Hu, Yulong Chen, Cheng-Lin Liu, Xu-Yao Zhang
arXiv:2602. 11908v3 Announce Type: replace Abstract: LLMs are widely used, yet they remain prone to factual errors that erode user trust and limit adoption in high-risk settings.
By Shani Goren, Ido Galil, Ran El-Yaniv
Large language models (LLMs) frequently produce confident yet factually incorrect responses when user inputs contain misleading premises, a phenomenon we attribute to fact perturbations in the input....
The paper introduces Evidence-Aligned Entity Verification (EAEV), a method for detecting entity-level hallucinations in retrieval-augmented generation (RAG). EAEV aligns generated entities with retrieved evidence across three dimensions and uses counterfactual stability analysis to maintain robust alignments when evidence changes. Experiments on multiple RAG benchmarks show that EAEV consistently outperforms existing hallucination detection methods and generalizes well.
By Runsong Jia, Zhen Fang, Mengjia Wu, Jie Lu, Yi Zhang