Verifiable by Design: Aligning Language Models to Quote from Pre-Training Data

arXiv:2404.03862v1 Announce Type: new
Abstract: For humans to trust the fluent generations of large language models (LLMs), they must be able to verify their correctness against trusted, external sources. Recent efforts aim to increase verifiability through citations of retrieved documents or post-hoc provenance. However, such citations are prone to mistakes that further complicate their verifiability. To address these limitations, we tackle the verifiability goal with a different philosophy: we trivialize the verification process by developing models that quote verbatim statements …

abstract aim arxiv citations cs.cl data design documents however humans language language models large language large language models llms pre-training provenance through training training data trust type verify