Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Huo, Nan, Li, Jinyang, Qin, Bowen, Qu, Ge, Li, Xiaolong, Li, Xiaodong, Ma, Chenhao, Cheng, Reynold
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2506.05278
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908574804869120
author Huo, Nan
Li, Jinyang
Qin, Bowen
Qu, Ge
Li, Xiaolong
Li, Xiaodong
Ma, Chenhao
Cheng, Reynold
author_facet Huo, Nan
Li, Jinyang
Qin, Bowen
Qu, Ge
Li, Xiaolong
Li, Xiaodong
Ma, Chenhao
Cheng, Reynold
contents Retrieval-Augmented Generation (RAG) systems commonly suffer from Knowledge Conflicts, where retrieved external knowledge contradicts the inherent, parametric knowledge of large language models (LLMs). It adversely affects performance on downstream tasks such as question answering (QA). Existing approaches often attempt to mitigate conflicts by directly comparing two knowledge sources in a side-by-side manner, but this can overwhelm LLMs with extraneous or lengthy contexts, ultimately hindering their ability to identify and mitigate inconsistencies. To address this issue, we propose Micro-Act a framework with a hierarchical action space that automatically perceives context complexity and adaptively decomposes each knowledge source into a sequence of fine-grained comparisons. These comparisons are represented as actionable steps, enabling reasoning beyond the superficial context. Through extensive experiments on five benchmark datasets, Micro-Act consistently achieves significant increase in QA accuracy over state-of-the-art baselines across all 5 datasets and 3 conflict types, especially in temporal and semantic types where all baselines fail significantly. More importantly, Micro-Act exhibits robust performance on non-conflict questions simultaneously, highlighting its practical value in real-world RAG applications.
format Preprint
id arxiv_https___arxiv_org_abs_2506_05278
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Micro-Act: Mitigating Knowledge Conflict in LLM-based RAG via Actionable Self-Reasoning
Huo, Nan
Li, Jinyang
Qin, Bowen
Qu, Ge
Li, Xiaolong
Li, Xiaodong
Ma, Chenhao
Cheng, Reynold
Computation and Language
Artificial Intelligence
Retrieval-Augmented Generation (RAG) systems commonly suffer from Knowledge Conflicts, where retrieved external knowledge contradicts the inherent, parametric knowledge of large language models (LLMs). It adversely affects performance on downstream tasks such as question answering (QA). Existing approaches often attempt to mitigate conflicts by directly comparing two knowledge sources in a side-by-side manner, but this can overwhelm LLMs with extraneous or lengthy contexts, ultimately hindering their ability to identify and mitigate inconsistencies. To address this issue, we propose Micro-Act a framework with a hierarchical action space that automatically perceives context complexity and adaptively decomposes each knowledge source into a sequence of fine-grained comparisons. These comparisons are represented as actionable steps, enabling reasoning beyond the superficial context. Through extensive experiments on five benchmark datasets, Micro-Act consistently achieves significant increase in QA accuracy over state-of-the-art baselines across all 5 datasets and 3 conflict types, especially in temporal and semantic types where all baselines fail significantly. More importantly, Micro-Act exhibits robust performance on non-conflict questions simultaneously, highlighting its practical value in real-world RAG applications.
title Micro-Act: Mitigating Knowledge Conflict in LLM-based RAG via Actionable Self-Reasoning
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2506.05278