Memorizing is Not Enough: Deep Knowledge Injection Through Reasoning

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Xu, Ruoxi, Ji, Yunjie, Cao, Boxi, Lu, Yaojie, Lin, Hongyu, Han, Xianpei, He, Ben, Sun, Yingfei, Li, Xiangang, Sun, Le
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912302391885824
author Xu, Ruoxi
Ji, Yunjie
Cao, Boxi
Lu, Yaojie
Lin, Hongyu
Han, Xianpei
He, Ben
Sun, Yingfei
Li, Xiangang
Sun, Le
author_facet Xu, Ruoxi
Ji, Yunjie
Cao, Boxi
Lu, Yaojie
Lin, Hongyu
Han, Xianpei
He, Ben
Sun, Yingfei
Li, Xiangang
Sun, Le
contents Although large language models (LLMs) excel in knowledge recall and reasoning, their static nature leads to outdated information as the real world evolves or when adapting to domain-specific knowledge, highlighting the need for effective knowledge injection. However, current research on knowledge injection remains superficial, mainly focusing on knowledge memorization and retrieval. This paper proposes a four-tier knowledge injection framework that systematically defines the levels of knowledge injection: memorization, retrieval, reasoning, and association. Based on this framework, we introduce DeepKnowledge, a synthetic experimental testbed designed for fine-grained evaluation of the depth of knowledge injection across three knowledge types (novel, incremental, and updated). We then explore various knowledge injection scenarios and evaluate the depth of knowledge injection for each scenario on the benchmark. Experimental results reveal key factors to reach each level of knowledge injection for LLMs and establish a mapping between the levels of knowledge injection and the corresponding suitable injection methods, aiming to provide a comprehensive approach for efficient knowledge injection across various levels.
format Preprint
id arxiv_https___arxiv_org_abs_2504_00472
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Memorizing is Not Enough: Deep Knowledge Injection Through Reasoning
Xu, Ruoxi
Ji, Yunjie
Cao, Boxi
Lu, Yaojie
Lin, Hongyu
Han, Xianpei
He, Ben
Sun, Yingfei
Li, Xiangang
Sun, Le
Computation and Language
Artificial Intelligence
Although large language models (LLMs) excel in knowledge recall and reasoning, their static nature leads to outdated information as the real world evolves or when adapting to domain-specific knowledge, highlighting the need for effective knowledge injection. However, current research on knowledge injection remains superficial, mainly focusing on knowledge memorization and retrieval. This paper proposes a four-tier knowledge injection framework that systematically defines the levels of knowledge injection: memorization, retrieval, reasoning, and association. Based on this framework, we introduce DeepKnowledge, a synthetic experimental testbed designed for fine-grained evaluation of the depth of knowledge injection across three knowledge types (novel, incremental, and updated). We then explore various knowledge injection scenarios and evaluate the depth of knowledge injection for each scenario on the benchmark. Experimental results reveal key factors to reach each level of knowledge injection for LLMs and establish a mapping between the levels of knowledge injection and the corresponding suitable injection methods, aiming to provide a comprehensive approach for efficient knowledge injection across various levels.
title Memorizing is Not Enough: Deep Knowledge Injection Through Reasoning
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2504.00472