<ul data-eligibleForWebStory="true"><li>A method is presented for validating RDF triples using LLMs with traceable arguments.</li><li>The approach avoids using internal LLM factual knowledge and instead compares verified RDF statements to external documents.</li><li>1,719 positive statements from the BioRED dataset were evaluated alongside the same number of newly generated negative statements, resulting in 88% precision and 44% recall, indicating the need for human oversight.</li><li>The method was also tested on the SNLI dataset, showing comparison with models tuned for natural language inference task.</li><li>The method was demonstrated on Wikidata using a SPARQL query to automatically retrieve statements for verification.</li><li>Results suggest that LLMs could be applied for large-scale validation of statements in knowledge graphs, reducing human annotation costs.</li></ul>

Traceable LLM-based validation of statements in knowledge graphs

Discover more