Retrieval-Augmented Generation for Large Language Models: A Survey
By Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jin Pan, Yuxi Bi, Yi Dai, Jiawei Sun, Qianyu Guo, Meng Wang and 1 more. 2023
3,697 citations recorded by the source corpus.
Large Language Models (LLMs) showcase impressive capabilities but encounter challenges like hallucination, outdated knowledge, and non-transparent, untraceable reasoning processes. Retrieval-Augmented Generation (RAG) has emerged as a promising solution by incorporating knowledge from external databases. This enhances the accuracy and credibility of the generation, particularly for knowledge-intensive tasks, and allows for continuous knowledge updates and integration of domain-specific information. RAG synergistically merges LLMs' intrinsic knowledge with the vast, dynamic repositories of external databases. This comprehensive review paper offers a detailed examination of the progression of RAG paradigms, encompassing the Naive RAG, the Advanced RAG, and the Modular RAG. It meticulously scrutinizes the tripartite foundation of RAG frameworks, which includes the retrieval, the generation and the augmentation techniques. The paper highlights the state-of-the-art technologies embedded in each of these critical components, providing a profound understanding of the advancements in RAG systems. Furthermore, this paper introduces up-to-date evaluation framework and benchmark. At the end, this article delineates the challenges currently faced and points out prospective avenues for research and development.
Subfields: Artificial Intelligence, Retrieval-Augmented Generation
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Kalpesh Krishna
- Manaal Faruqui
- Dawei Yin
- Yoshua Bengio
- Thomas Laurent
- Zhijiang Guo
- Nitesh V. Chawla
- Cunxiang Wang
In Retrieval-Augmented Generation
Citations in the corpus
- Fact, Fetch, and Reason: A Unified Evaluation of Retrieval-Augmented Generation (2024, background): “These techniques are particularly effective for tasks requiring multi-hop reasoning, factual grounding, and synthesizing information from diverse knowledge domains (Gao et al., 2023).”
- Rethinking machine unlearning for large language models (2024, neutral): “Unlearning can be a natural solution for both types of attacks given the existing success of unlearning as a defense against adversarial attacks in other domains (Wang et al., 2019; Li et al., 2021; Liu et al., 2022b; Jia et al., 2023).”
- Retrieval-Augmented Generation for AI-Generated Content: A Survey (2024, neutral): “Most of the existing works focus on text-related RAG tasks that are facilitated by LLMs, without in-depth investigation in other modalities.”
- Knowledge Conflicts for LLMs: A Survey (2024, background): “However, intra-memory conflicts, where LLMs generate differing responses to similar inputs, undermine their reliability and utility by introducing undesirable uncertainty.”
- Large language models for generative information extraction: a survey (2023, positive): “Li et al. [161] addressed the challenge of document-level relation extraction from a long context, and proposed an automated annotation method for DocRE that combines a LLM with a natural language inference module to generate relation triples.”
- Benchmarking Retrieval-Augmented Generation for Medicine (2024, background): “Many follow-up studies have been carried out to improve over the vanilla RAG (Borgeaud et al., 2022; Ram et al., 2023; Gao et al., 2023; Jiang et al., 2023; Mialon et al., 2023).”
- Tool learning with large language models: a survey (2024, positive): “Meanwhile, we also acknowledge the foundational contributions of earlier perspective papers like those by Mialon et al. (2023) [45] and Qin et al. (2023) [16], which initially highlighted the promising opportunities that tools present to enhance LLMs capabilities.”
- mGTE: Generalized Long-Context Text Representation and Reranking Models for Multilingual Text Retrieval (2024, background): “Recent advances in large language models (LLMs) and retrieval augmented generation (RAG) (Gao et al., 2023) systems have led to an unprecedented surge in demand for versatile, plug-and-play TRMs and rerankers.”
- Large Language Model based Multi-Agents: A Survey of Progress and Challenges (2024, methodology): “Building on this, [Gao et al. , 2023a] takes the concept further by constructing vast networks comprising 8,563 and 17,945 agents, respectively, designed to simulate social networks focused on the topics of Gender Discrimination and Nuclear Energy.”
- From Local to Global: A Graph RAG Approach to Query-Focused Summarization (2024, methodology): “Na ¨ ıve RAG approaches (Gao et al., 2023) do this by converting documents to text, splitting text into chunks, and embedding these chunks into a vector space in which similar positions represent similar semantics.”
- Graph Retrieval-Augmented Generation: A Survey (2024, neutral): “How did the artistic movements of the 19th century impact the development of modern art in the 20th century?”
- LightRAG: Simple and Fast Retrieval-Augmented Generation (2024, neutral): “Complex Query Handling : This approach is especially valuable in scenarios requiring diverse perspectives.”
- G-Retriever: Retrieval-Augmented Generation for Textual Graph Understanding and Question Answering (2024, positive): “The concept of Retrieval-Augmented Generation, initially proposed by Lewis et al. [21], has gained increased attention for its ability to mitigate the issue of hallucination within LLMs and enhance trustworthiness and explainability [7].”
- A-MEM: Agentic Memory for LLM Agents (2025, neutral): “Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains, including natural language processing, code generation, and recommender systems (Wang et al., 2023b; Zhang et al., 2024a; Xu et al., 2024a,b, 2023).”
- Large Language Models for Education: A survey and outlook (2024, positive): “Additionally, RAG has proved to be an effective method for addressing hallucinations and inconsistencies, thereby improving the reliability and accuracy of LLM-generated content [33].”
- Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning (2025, positive): “• We conduct systematic experiments to demonstrate the effectiveness of S EARCH -R1, with two LLMs achieving respective average relative improvements of 41% and 20% over RAG baselines under the same experimental setup ( e.g. , same retrieval model, training data, and pre-trained LLMs).”