通过双流知识融合，有效减少大模型幻觉。
Bridging External and Parametric Knowledge: Mitigating Hallucination of LLMs with Shared-Private Semantic Synergy in Dual-Stream Knowledge
- 设计双流架构，区分共享与私有语义，控制知识引入
- 在多个数据集上优于基线模型，显著降低幻觉率
- 适合需要高可信生成的场景，如医疗、法律问答
检索增强生成（RAG）通过引入外部知识来缓解大语言模型（LLM）的幻觉问题。然而，外部知识可能包含噪声或与模型内部参数知识冲突，导致性能下降。当前的LLM缺乏解决此类冲突的内在机制。为此，我们提出双流知识增强框架DSSP-RAG，其核心是将传统自注意力改进为混合注意力，以区分共享与私有语义，实现可控的知识融合。同时，提出一种无监督幻觉检测方法，基于模型内在认知不确定性判断是否引入外部知识。为减少外部知识噪声，引入能量比（EQ），通过任务对齐与任务错位层间的注意力差异矩阵进行度量。大量实验表明，DSSP-RAG在多个基准测试中均优于强基线模型。
原文摘要 · Abstract (English)
Retrieval-augmented generation (RAG) aims to mitigate the hallucination of Large Language Models (LLMs) by retrieving and incorporating relevant external knowledge into the generation process. However, the external knowledge may contain noise and conflict with the parametric knowledge of LLMs, leading to degraded performance. Current LLMs lack inherent mechanisms for resolving such conflicts. To fill this gap, we propose a Dual-Stream Knowledge-Augmented Framework for Shared-Private Semantic Synergy (DSSP-RAG). Central to it is the refinement of the traditional self-attention into a mixed-attention that distinguishes shared and private semantics for a controlled knowledge integration. An unsupervised hallucination detection method that captures the LLMs' intrinsic cognitive uncertainty ensures that external knowledge is introduced only when necessary. To reduce noise in external knowledge, an Energy Quotient (EQ), defined by attention difference matrices between task-aligned and task-misaligned layers, is proposed. Extensive experiments show that DSSP-RAG achieves a superior performance over strong baselines.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。