Quantifying reliance on external information over parametric knowledge during Retrieval Augmented Generation (RAG) using mechanistic analysis
원문 초록
Retrieval Augmented Generation (RAG) is a widely used approach for leveraging external context in several natural language applications such as question answering and information retrieval. Yet, the exact nature in which a Language Model (LM) leverages this non-parametric memory or retrieved context isn't clearly understood. This paper mechanistically examines the RAG pipeline to highlight that LMs demonstrate a "shortcut'' effect and have a strong bias towards utilizing the retrieved context to answer questions, while relying minimally on model priors. We propose (a) Causal Mediation Analysis; for proving that parametric memory is minimally utilized when answering a question and (b) Attention Contributions and Knockouts for showing the last token residual stream do not get enriched from the subject token in the question, but gets enriched from tokens of RAG-context. We find this pronounced "shortcut'' behaviour to be true across both LLMs (e.g.,LlaMa) and SLMs (e.g., Phi)
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
From RAG to Agentic: Validating Islamic-Medicine Responses with LLM Agents
Describe Anything Model for Visual Question Answering on Text-rich Images
Lightweight Relevance Grader in RAG
SeqPE: Transformer with Sequential Position Encoding
XGraphRAG: Interactive Visual Analysis for Graph-based Retrieval-Augmented Generation
Is This Just Fantasy? Language Model Representations Reflect Human Judgments of Event Plausibility
토론
이 논문과 연결된 포럼 글이 아직 없습니다.