Abstract

Methods for relation extraction from text mostly focus on high precision, atthe cost of limited recall. High recall is crucial, though, to populate longlists of object entities that stand in a specific relation with a givensubject. Cues for relevant objects can be spread across many passages in longtexts. This poses the challenge of extracting long lists from long texts. Wepresent the L3X method which tackles the problem in two stages: (1)recall-oriented generation using a large language model (LLM) with judicioustechniques for retrieval augmentation, and (2) precision-orientedscrutinization to validate or prune candidates. Our L3X method outperformsLLM-only generations by a substantial margin.