Featured · Paper · itc-r-002
Isnad-style provenance for retrieval-augmented generation
Abstract
Language models that cite sources often cite them wrongly, and the reader has no easy way to check. The classical hadith sciences faced the same problem and solved it with the isnād, a chain of transmission that names every link between a report and its origin. We propose an analogous structure for retrieval-augmented generation, so that every claim carries a chain back to a primary source.
Placeholder abstract. Replace with the published version.