RECoT: Relation-enhanced Chains-of-Thoughts for knowledge-intensive multi-hop questions answering

Download PDF
Open Webpage

Ronghan Li, Dongdong Li, Haowen Yang, Xiaoxi Liu, Haoxiang Jin, Rongcheng Pu, Qiguang Miao

Published: 2025, Last Modified: 04 Jan 2026Neurocomputing 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•This paper proposes a method for question decomposition guided by relational triples.•This paper introduces to select the optimal CoT reasoning sentence based on relational triples.•RECoT improves retrieval and QA performance on prevalent open-domain QA datasets.