SCoNE: Selective Context-aware Neuron Editing for Robust Retrieval-Augmented Generation

2026-09-01

AI Summary

为解决检索增强生成中因检索噪音导致的模型易分心问题，提出了一种无需训练的选择性上下文感知神经元编辑方法SCoNE，通过强化特定神经元来提高鲁棒性。

Abstract

Retrieval-Augmented Generation (RAG) is highly sensitive to retrieval noise: when retrieved documents mix informative and irrelevant context, LLMs are easily distracted, leading to hallucinations. To overcome this, we propose SCoNE (Selective Context-aware Neuron Editing), a training-free model editing approach that improves retrieval noise robustness by selectively strengthening context-aware FFN neurons that are identified by both high attribution and high cross-input variability. SCoNE requires only a small number of mining samples, no fine-tuning, and no inference-time overhead. Across various knowledge-intensive question-answering benchmarks and two LLM backbones, SCoNE consistently outperforms competitive baseline methods. Our code is available at https://github.com/HYU-ARK-Lab/SCoNE.

Problem

• Retrieval-Augmented Generation
• retrieval noise
• hallucinations

Innovation

• Selective Context-aware Neuron Editing
• Retrieval Noise Robustness
• Training-free Model Editing