Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Jin, Jiajie, Zhu, Yutao, Zhou, Yujia, Dou, Zhicheng
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2402.12174
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909213066788864
author Jin, Jiajie
Zhu, Yutao
Zhou, Yujia
Dou, Zhicheng
author_facet Jin, Jiajie
Zhu, Yutao
Zhou, Yujia
Dou, Zhicheng
contents Retrieval-augmented large language models (LLMs) have demonstrated efficacy in knowledge-intensive tasks such as open-domain QA, addressing inherent challenges in knowledge update and factual inadequacy. However, inconsistencies between retrieval knowledge and the necessary knowledge for LLMs, leading to a decline in LLM's answer quality. This paper introduces BIDER, an approach that refines retrieval documents into Key Supporting Evidence (KSE) through knowledge synthesis, supervised fine-tuning (SFT), and preference alignment. We train BIDER by learning from crafting KSE, while maximizing its output to align with LLM's information acquisition preferences through reinforcement learning. Evaluations across five datasets show BIDER boosts LLMs' answer quality by 7% while reducing input content length in retrieval documents by 80%, outperforming existing methods. The proposed KSE simulation effectively equips LLMs with essential information for accurate question answering.
format Preprint
id arxiv_https___arxiv_org_abs_2402_12174
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle BIDER: Bridging Knowledge Inconsistency for Efficient Retrieval-Augmented LLMs via Key Supporting Evidence
Jin, Jiajie
Zhu, Yutao
Zhou, Yujia
Dou, Zhicheng
Computation and Language
Retrieval-augmented large language models (LLMs) have demonstrated efficacy in knowledge-intensive tasks such as open-domain QA, addressing inherent challenges in knowledge update and factual inadequacy. However, inconsistencies between retrieval knowledge and the necessary knowledge for LLMs, leading to a decline in LLM's answer quality. This paper introduces BIDER, an approach that refines retrieval documents into Key Supporting Evidence (KSE) through knowledge synthesis, supervised fine-tuning (SFT), and preference alignment. We train BIDER by learning from crafting KSE, while maximizing its output to align with LLM's information acquisition preferences through reinforcement learning. Evaluations across five datasets show BIDER boosts LLMs' answer quality by 7% while reducing input content length in retrieval documents by 80%, outperforming existing methods. The proposed KSE simulation effectively equips LLMs with essential information for accurate question answering.
title BIDER: Bridging Knowledge Inconsistency for Efficient Retrieval-Augmented LLMs via Key Supporting Evidence
topic Computation and Language
url https://arxiv.org/abs/2402.12174