04:00
2026-09-10
arxiv.org
large-language-models
SWORD: Wikidata-based Distortions Reveal Hidden Cross-Lingual Inconsistencies in LLM Factual Error Rejection
A new arXiv paper introduces SWORD (Systematic Wikidata-based Object-Relation Distortion), a benchmark that tests whether large language models consistently reject factually incorrect statements acros…