Awesome-Graph4LLM

This repository contains a curated list of resources on graph-enhanced large language models (Graph4LLM).

The link to the paper is as follows: Are Graphs Useful for LLMs? A Comprehensive Survey of Graph-Enhanced Large Language Models

🎉News

[2026-04-30] Our paper Graph4LLM: A Systematic Survey of Graph-Enhanced Large Language Models is accepted by IJCAI 2026.

Overview

This paper presents a systematic and pipeline-oriented survey of Graph4LLM. We categorize existing Graph4LLM methods according to the three core phases of the LLM pipeline: input phase, where graphs structure prompts and incorporate external knowledge; model phase, where graphs guide internal computation or multi-agent coordination; and output phase, where graphs support structured reasoning, planning and verification. For each phase, we conduct a detailed review of key methods and their design trade-offs, further elaborate on the broad application scenarios of Graph4LLM in general tasks and specific domains, and finally outline the critical challenges and promising future research directions for this field.

Research Papers

📑 Table of Contents

Awesome-Graph4LLM

🔹 Input Phase

Prompt from Knowledge Graph

Graph-Enhanced Instruction Tuning

(ACL 2024) ChatKBQA: A Generate-then-Retrieve Framework for Knowledge Base Question Answering [Paper]
(CIKM 2024) A GAIL Fine-Tuned LLM Enhanced Framework for Low-Resource Knowledge Graph Question Answering [Paper]
(AAAI-MAKE 2024) GLaM: Fine-Tuning Large Language Models for Domain Knowledge Graph Alignment via Neighborhood Partitioning and Generative Subgraph Encoding [Paper]
(ICLR 2025) Knowledge Graph Finetuning Enhances Knowledge Manipulation in Large Language Models [Paper]
(ESWA 2025) Exploiting knowledge graph communities to fine-tune large language models [Paper]

Graph-Retrieval Augmented Generation

(NLRSE 2023) Knowledge-Augmented Language Model Prompting for Zero-Shot Knowledge Graph Question Answering [Paper]
(EMNLP 2023) StructGPT: A General Framework for Large Language Model to Reason over Structured Data [Paper]
(ACL 2024) MindMap: Knowledge Graph Prompting Sparks Graph of Thoughts in Large Language Models [Paper]
(ICLR 2025) Simple is Effective: The Roles of Graphs and Large Language Models in Knowledge-Graph-Based Retrieval-Augmented Generation [Paper]
(ICLR 2024) Think-on-Graph: Deep and Responsible Reasoning of Large Language Model on Knowledge Graph [Paper]
(ICLR 2024) Reasoning on Graphs: Faithful and Interpretable Large Language Model Reasoning [Paper]
(AAAI 2024) Mitigating Large Language Model Hallucinations via Autonomous Knowledge Graph-Based Retrofitting [Paper]
(ACL 2025) GNN-RAG: Graph Neural Retrieval for Large Language Model Reasoning [Paper]

Prompt from Corpus-Level Graph

(ArXiv 2024) From Local to Global: A GraphRAG Approach to Query-Focused Summarization [Paper]
(ICLR 2024) RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval (RAPTOR) [Paper]
(EMNLP 2025) LightRAG: Simple and Fast Retrieval-Augmented Generation [Paper]
(NeurIPS 2024) G-Retriever: Retrieval-Augmented Generation for Textual Graph Understanding and Question Answering [Paper]
(NeurIPS 2024) HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models [Paper]
(ICML 2025) HippoRAG2: From RAG to Memory: Non-Parametric Continual Learning for Large Language Models [Paper]
(NeurIPS 2025) GFM-RAG: Graph Foundation Model for Retrieval Augmented Generation [Paper]
(AAAI 2026) PathRAG: Pruning Graph-Based Retrieval Augmented Generation with Relational Paths [Paper]
(ICLR 2026) LinearRAG: Linear Graph Retrieval Augmented Generation on Large-scale Corpora [Paper]

Prompt from Instance-Level Graph

(ACL 2024) GraphReader: Building Graph‑based Agent to Enhance Long‑Context Abilities of Large Language Models [Paper]
(ACL 2024) Large Language Models Can Learn Temporal Reasoning [Paper]
(ACL 2025) Reasoning with Graphs: Structuring Implicit Knowledge to Enhance LLMs Reasoning [Paper]
(PAKDD 2025) GRL‑Prompt: Towards Knowledge Graph based Prompt Optimization via Reinforcement Learning [Paper]
(EMNLP 2024) Structure Guided Prompt: Instructing Large Language Model in Multi‑Step Reasoning by Exploring Graph Structure of the Text [Paper]

🔹 Model Phase

Single Model

External Graph Adapter

(MM 2023) Making Large Language Models Perform Better in Knowledge Graph Completion [Paper]
(COLING 2025) Filter‑then‑Generate: Large Language Models with Structure‑Text Adapter for Knowledge Graph Completion [Paper]
(EMNLP 2024) GraphAdapter: Tuning Vision‑Language Models With Dual Knowledge Graph [Paper]
(EMNLP 2023) MolCA: Molecular Graph‑Language Modeling with Cross‑Modal Projector and Uni‑Modal Adapter [Paper]
(ICLR 2026) Graph Tokenization for Bridging Graphs and Transformers [Paper]

Internal Model Fusion

(ICLR 2022) GreaseLM: Graph REASoning Enhanced Language Models [Paper]
(NeurIPS 2021) GraphFormers: GNN‑nested Transformers for Representation Learning on Textual Graph [Paper]
(EMNLP 2022) Graph‑Induced Transformers for Efficient Multi‑Hop Question Answering[Paper]
(ACL 2024) GNNAVI: Navigating the Information Flow in Large Language Models by Graph Neural Network [Paper]
(ACL 2024 Findings) KG‑Adapter: Enabling Knowledge Graph Integration in Large Language Models through Parameter‑Efficient Fine‑Tuning (KG‑Adapter) [Paper]
(ACL 2023) Patton: Language Model Pretraining on Text‑Rich Networks (Patton) [Paper]

Multiple Models

Static Topology

(LOG 2025) FLORA: GNNs as Predictors of Agentic Workflow Performances [Paper]
(ICLR 2025) Scaling Large Language Model‑based Multi‑Agent Collaboration [Paper]
(NeurIPS 2023) CAMEL: Communicative Agents for “Mind” Exploration of Large Language Model Society [Paper]
(AAAI 2026) S‑DAG: A Subject‑Based Directed Acyclic Graph for Multi‑Agent Heterogeneous Reasoning [Paper]
(AAAI 2024) TAPE: Leveraging Agent Topology for Cooperative Multi‑Agent Policy Gradient [Paper]

Dynamic Topology

(ICLR 2025) Cut the Crap: An Economical Communication Pipeline for LLM‑based Multi‑Agent Systems [Paper]
(ICML 2024) GPTSwarm: Language Agents as Optimizable Graphs [Paper]
(EMNLP 2025) AMAS: Adaptively Determining Communication Topology for LLM‑based Multi‑Agent System [Paper]
(COLM 2024) A Dynamic LLM‑Powered Agent Network for Task‑Oriented Agent Collaboration [Paper]
(ICLR 2026) Graph‑of‑Agents: A Graph‑based Framework for Multi‑Agent LLM Collaboration [Paper]
(ICLR 2026) CARD: Towards Conditional Design of Multi‑Agent Topological Structures [Paper]
(NeurIPS 2025) AgentNet: Decentralized Evolutionary Coordination for LLM‑based Multi‑Agent Systems [Paper]

🔹 Output Phase

Reasoning

Reasoning Based on Node Summarization

(NACCL 2025) THOUGHTSCULPT: Reasoning with Intermediate Revision and Search [Paper]
(AAAI 2025) RATT: A Thought Structure for Coherent and Correct LLM Reasoning [Paper]
(AAAI 2024) Graph of Thoughts: Solving Elaborate Problems with Large Language Models [Paper]
(BigData 2024) Graph of Logic: Enhancing LLM Reasoning with Graphs and Symbolic Logic [Paper]
(TMLR 2025) Cumulative Reasoning with Large Language Models [Paper]

Reasoning Based on Path Exploration

(NeurIPS 2023) Tree of Thoughts: Deliberate Problem Solving with Large Language Models [Paper]
(ICLR 2024) Skeleton‑of‑Thought: Prompting LLMs for Efficient Parallel Generation [Paper]
(ICML 2024) Language Agent Tree Search Unifies Reasoning, Acting, and Planning in Language Models [Paper]
(ACL 2024) Everything of Thoughts: Defying the Law of Penrose Triangle for Thought Generation [Paper]
(ICML 2025) Policy Guided Tree Search for Enhanced LLM Reasoning [Paper]
(ACL 2025 Findings) Self‑attention‑based Graph‑of‑Thought for Math Problem Solving [Paper]
(ICLR 2026) DAG‑Math: Graph‑Guided Mathematical Reasoning in LLMs [Paper]

Planning

(EMNLP 2023) Reasoning with Language Model is Planning with World Model [Paper]
(NeurIPS 2024) Can Graph Learning Improve Planning in LLM‑based Agents? [Paper]
(ECCV 2024) ControlLLM: Augment Language Models with Tools by Searching on Graphs [Paper]
(NeurIPS 2024) Plan‑on‑Graph: Self‑Correcting Adaptive Planning of Large Language Model on Knowledge Graphs [Paper]
(IJAIRR 2024) Consolidating Trees of Robotic Plans Generated Using Large Language Models to Improve Reliability [Paper]

Evaluation

Factuality Evaluation

(NAACL 2025) A Probabilistic Framework for LLM Hallucination Detection via Belief Tree Propagation [Paper]
(AAAI 2025) Enhancing Uncertainty Modeling with Semantic Graph for Hallucination Detection [Paper]
(ArXiv 2024) Graph‑based Uncertainty Metrics for Long‑form Language Model Outputs [Paper]
(ArXiv 2024) Graph‑based Confidence Calibration for Large Language Models [Paper]
(KiL 2024) GraphEval: A Knowledge‑Graph Based LLM Hallucination Evaluation Framework [Paper]
(TrustNLP 2024) FactAlign: Fact‑Level Hallucination Detection and Classification Through Knowledge Graph Alignment [Paper]
(AAAI 2026) Assessing Automated Fact‑Checking for Medical LLM Responses with Knowledge Graphs [Paper]
(AAAI 2025) Zero‑resource Hallucination Detection for Text Generation via Graph‑based Contextual Knowledge Triples Modeling [Paper]
(EACL 2026) FactSelfCheck: Fact‑Level Black‑Box Hallucination Detection for LLMs [Paper]

Logicality Evaluation

(ACL 2025) Semantic‑Eval: A Semantic Comprehension Evaluation Framework for Large Language Models Generation without Training [Paper]
(EMNLP 2025) DiagramEval: Evaluating LLM‑Generated Diagrams via Graphs [Paper]
(ACL 2024) GraphReason: Enhancing Reasoning Capabilities of Large Language Models through a Graph‑Based Verification Approach [Paper]
(EMNLP 2025) Mapping the Minds of LLMs: A Graph‑Based Analysis of Reasoning LLMs [Paper]
(NeurIPS 2025) Topology of Reasoning: Understanding Large Reasoning Models through Reasoning Graph Properties [Paper]

🚀 Applications of Graph4LLM

🔹 General Applications

Classic NLP

(ISWC 2024) Knowledge Graphs for Enhancing Large Language Models in Entity Disambiguation [Paper]
(CCKS-IJCKG 2024) An LLM-SPARQL Hybrid Framework for Named Entity Linking and Disambiguation to Wikidata [Paper]
(ACL 2024 (Findings)) GNNAVI: Navigating the Information Flow in Large Language Models by Graph Neural Network [Paper]
(AIMSA 2024) Combining Graph NN and LLM for Improved Text-Based Emotion Recognition [Paper]
(KSEM 2025) SKG-LLM: Enhancing Large Language Models with Sentiment Knowledge Graphs for Fine-Grained Sentiment Analysis [Paper]
(ACL 2023) GNN-SL: Sequence Labeling Based on Nearest Examples via GNN [Paper]
(ACL 2022) Cross-domain Named Entity Recognition via Graph Matching [Paper]
(ECCC 2025) LLM-assisted Graph-RAG Information Extraction from IFC Data [Paper]
(ACL 2025 (Findings)) GLiM: Integrating Graph Transformer and LLM for Document-Level Biomedical Relation Extraction with Incomplete Labeling [Paper]
(CIKM 2025) RePooL: Relation-Faceted Graph Pooling with LLM Guidance for Dynamic Span-Aware Information Extraction [Paper]
(EMNLP 2025) GDLLM: A Global Distance-aware Modeling Approach Based on Large Language Models for Event Temporal Relation Extraction [Paper]
(EMNLP 2025) GRAFT: A Graph-based Flow-aware Agentic Framework for Document-level Machine Translation [Paper]
(ACL 2024) POMP: Probability-driven Meta-graph Prompter for LLMs in Low-resource Unsupervised Neural Machine Translation [Paper]
(ACL 2025) Graph of Records: Boosting Retrieval Augmented Generation for Long-context Summarization with Graphs [Paper]
(TCDS 2024) Adaptive Multidocument Summarization Via Graph Representation Learning [Paper]

Code

(TSEM 2025) Structured Chain-of-Thought Prompting for Code Generation [Paper]
(JSS 2024) GRACE: Empowering LLM-based software vulnerability detection with graph structure and in-context learning [Paper]
(ACL 2025) GALLa: Graph Aligned Large Language Models for Improved Source Code Understanding [Paper]
(FSE 2024) CodePlan: Repository-Level Coding using LLMs and Planning [Paper]
(ICLR 2025) RepoGraph: Enhancing AI Software Engineering with Repository-level Code Graph [Paper]
(FORGE 2025) RepoHyper: Search-Expand-Refine on Semantic Graphs for Repository-Level Code Completion [Paper]
(ACL 2025) CodexGraph: Bridging Large Language Models and Code Repositories via Code Graph Databases [Paper]
(ACL 2025) LocAgent: Graph-Guided LLM Agents for Code Localization [Paper]
(NeurIPS 2025) Code Graph Model (CGM): A Graph-Integrated Large Language Model for Repository-Level Software Engineering Tasks [Paper]

Table

(AAAI 2025) HeGTa: Leveraging Heterogeneous Graph-enhanced Large Language Models for Few-shot Complex Table Understanding [Paper]
(AAAI 2025) TabGLM: Tabular Graph Language Model for Learning Transferable Representations Through Multi-Modal Consistency Minimization [Paper]
(TRL @ NeurIPS 2024) RACOON: An LLM-based Framework for Retrieval-Augmented Column Type Annotation with a Knowledge Graph [Paper]
(COLING 2025) GraphOTTER: Evolving LLM-based Graph Reasoning for Complex Table Question Answering [Paper]
(NAACL 2025) Hybrid Graphs for Table-and-Text based Question Answering using LLMs [Paper]
(EMNLP 2025) Plugging schema graph into multi-table qa: A human-guided framework for reducing llm reliance [Paper]
(COLM 2025) Improving Table Understanding with LLMs and EntityOriented Search [Paper]
(ISWC 2025) KGSynX: Knowledge Graph and Explainable Feedback Guided LLMs for Synthetic Tabular Data Generation [Paper]

🔹 Domain-Specific Applications

E-Commerce

(EMNLP 2024) Enhancing High-order Interaction Awareness in LLM-based Recommender Model [Paper]
(WSDM 2025, ACM International Conference on Web Search and Data Mining) Enhancing Collaborative Semantics of Language Model-Driven Recommendations via Graph-Aware Learning [Paper]
(WSDM 2024) LLMRec: Large Language Models with Graph Augmentation for Recommendation [Paper]
(KBS 2025) LLM-AGR: Large Language Model Augmented Graph Representation Learning for Recommendation [Paper]
(SIGIR 2025) CORONA: A Coarse-to-Fine Framework for Graph-based Recommendation with Large Language Models [Paper]
(ICDM 2025) Reasoning over User Preferences: Knowledge Graph-Augmented LLMs for Explainable Conversational Recommendations [Paper]
(AAAI 2025) Exploring large language model for graph data understanding in online job recommendations [Paper]

Medicine

(ACL 2025) Medical Graph RAG: Evidence-based Medical Large Language Model via Graph Retrieval-Augmented Generation [Paper]
(Bioinformatics 2024) Biomedical knowledge graph-optimized prompt generation for large language models [Paper]
(KAIS 2025) LLM-KGMQA: large language model-augmented multi-hop question-answering system based on knowledge graph in medical field [Paper]
(ICLR 2025) Multimodal Large Language Models for Inverse Molecular Design with Retrosynthetic Planning [Paper]
(Information Fusion 2026) Magic: AN LLM-based multi-agent activated graph-reasoning intelligent collaboration model for liver disease diagnosis [Paper]
(KDD 2024) Diet-ODIN: A Novel Framework for Opioid Misuse Detection with Interpretable Dietary Patterns [Paper]

Education

(EMOOCs 2025) Leveraging Graph Retrieval-Augmented Generation to Support Learners’ Understanding of Knowledge Concepts in MOOCs [Paper]
(AIQAM 2024) Cross-data knowledge graph construction for LLM-enabled educational question-answering system: A case study at HCMUT [Paper]
(MDPI 2025) Synergizing Knowledge Graphs and LLMs: An Intelligent Tutoring Model for Self-Directed Learning [Paper]
(CAAI 2024) FOKE: A Personalized and Explainable Education Framework Integrating Foundation Models, Knowledge Graphs, and Prompt Engineering [Paper]
(The Visual Computer 2025) MAS-KCL: knowledge component graph structure learning with large language model-based agentic workflow [Paper]
(ICCPCT 2025) Adaptive Machine Learning Frameworks Integrating LLMs and Knowledge Graphs for Tailored Education [Paper]

Finance

(ACL 2025) FinRipple: Aligning Large Language Models with Financial Market for Event Ripple Effect Awareness [Paper]
(ICAIF 2024) FinDKG: Dynamic Knowledge Graphs with Large Language Models for Detecting Global Trends in Financial Markets [Paper]
(GenAI in Finance 2025) FinCARE: Financial Causal Analysis with Reasoning and Evidence [Paper]
(GenAI-K 2025) GraphRAG: Leveraging Graph-Based Efficiency to Minimize Hallucinations in LLM-Driven RAG for Finance Data [Paper]
(ECML PKDD 2024) FinQA: A Training-Free Dynamic Knowledge Graph Question Answering System in Finance with LLM-Based Revision [Paper]
(DECS 2025) Construction of Financial Risk Assessment Model Based on Text Mining and LLM Architecture [Paper]
(SECA 2025) Research on Personalized Financial Product Recommendation by Integrating Large Language Models and Graph Neural Networks [Paper]

Law

(ICIC 2024) Legal-LM: Knowledge Graph Enhanced Large Language Models for Law Consulting [Paper]
(AIITA 2025s) Law GraphRAG: An Advanced Legal Question-Answering System [Paper]
(ACIIDS 2025) Knowledge Graph-Based Legal Query System with LLM and Retrieval Augmented Generation [Paper]
(CSAI 2024) Legal Judgment Prediction with LLM and Graph Contrastive Learning Networks (LLG-Judger) [Paper]
(MDPI 2024) Construction of Legal Knowledge Graph Based on Knowledge-Enhanced Large Language Models (JKEM) [Paper]
(Artificial Intelligence and Law 2025) LegisSearch: navigating legislation with graphs and large language models [Paper]

EDA

(DAC 2025) NetTAG: A Multimodal RTL-and-Layout-Aligned Netlist Foundation Model via Text-Attributed Graph [Paper]
(DAC 2025) CircuitFusion: Multimodal Circuit Representation Learning for Agile Chip Design [Paper]
(ICCAD 2024) Graph-Based Timing Prediction at Early-Stage RTL Using Large Language Model [Paper]
(DAC 2025) MOSS: Multi-Modal Representation Learning on Sequential Circuits [Paper]
(FPGA 2025) SAGE-HLS: Syntax-Aware AST-Guided LLM for High-Level Synthesis Code Generation [Paper]
(ASP-DAC 2025) RTL++: Graph-enhanced LLM for RTL Code Generation [Paper]
(ICCAD 2025) SynC-LLM: Generation of Large-Scale Synthetic Circuit Code with Hierarchical Language Models [Paper]
(DATE 2025) ASPEN: LLM-Guided E-Graph Rewriting for RTL Datapath Optimization [Paper]
(NeurIPS 2025) Evolving Graph Structured Programs for Circuit Generation with Large Language Models [Paper]
(ICML 2025) BRIDGES: Bridging Graph Modality and Large Language Models within EDA Tasks [Paper]

Survey Methodology

This section documents how the literature in this repository is collected, screened, verified, and classified. The goal is to make the survey process reproducible while allowing this repository to remain a living resource. The literature cutoff for the current repository snapshot is September 9, 2026. Papers added after a manuscript-specific cutoff are retained in the repository history and can be distinguished through Git commits.

Literature Sources

We search the following sources:

scholarly search and metadata services: Google Scholar, Semantic Scholar, DBLP, and arXiv;
NLP and machine-learning proceedings: ACL Anthology, OpenReview, NeurIPS Proceedings, and PMLR;
publisher and conference libraries: ACM Digital Library, IEEE Xplore, AAAI Proceedings, IJCAI Proceedings, SpringerLink, and relevant journal websites; and
backward and forward reference tracing from included surveys and representative Graph4LLM papers.

Whenever possible, publication metadata is verified against an official proceedings or publisher page. arXiv is used for discovery and for papers that do not yet have a confirmed peer-reviewed version.

Search Queries

The search combines terms describing large language models with terms describing graph structures and the role played by the graph. The principal Boolean query families are:

("large language model" OR LLM) AND
(graph OR "knowledge graph" OR GNN OR "graph neural network") AND
(augment OR enhance OR prompt OR retrieval OR reasoning OR planning OR evaluation)

(GraphRAG OR "graph retrieval-augmented generation" OR
 "knowledge graph prompting" OR "graph-enhanced prompting") AND
(LLM OR "large language model")

("large language model" OR "LLM agent") AND
("graph reasoning" OR "graph planning" OR "graph search" OR
 "graph verification" OR "agent graph" OR "communication graph")

("LLM agent" OR "language agent") AND
(skill OR tool OR workflow) AND
(graph OR dependency OR topology OR DAG)

Queries are also specialized for individual application domains by adding terms such as code, table, recommendation, medicine, education, finance, law, or electronic design automation. We additionally search exact method names and terminology variants identified during citation tracing.

Inclusion and Exclusion Criteria

A work is included when it satisfies all of the following conditions:

An LLM or LLM-based agent is a central component of the system.
A graph has a functional role in enhancing an LLM-centered process, rather than appearing only as a data format or visualization.
The graph affects at least one part of the LLM pipeline: input construction or retrieval, model-level computation or coordination, or output reasoning, planning, generation, or evaluation.
The work provides enough technical information to identify how the graph is constructed or obtained, how it interacts with the LLM, and what task the combined system performs.
A traceable scholarly record is available through official proceedings, a publisher page, or a public preprint repository.

We exclude works when the main objective is a conventional graph-learning or graph-prediction task and the LLM is used only as a text encoder, feature generator, or interface. Such work is treated as LLM4Graph rather than Graph4LLM. We also exclude papers in which graphs are mentioned only conceptually, used solely for post-hoc visualization, or do not influence the LLM workflow. Non-scholarly webpages and records whose title or publication identity cannot be verified are not included.

Earlier pretrained language models and multimodal language models are considered under the same functional criterion: they are included only when the graph directly enhances the language-model-centered pipeline covered by this survey.

Screening Procedure

Candidate papers are processed in the following stages:

Discovery: Retrieve candidates using the query families above, citation tracing, and exact-title searches.
Title and abstract screening: Remove records that are graph-related or LLM-related but do not study graphs enhancing an LLM-centered process.
Full-text eligibility check: Inspect the method, task definition, experiments, and graph–LLM interaction to confirm scope and determine the graph's primary function.
Metadata verification: Check the title, authors, year, venue, DOI or arXiv identifier, and paper URL against authoritative sources.
Taxonomy assignment: Record the primary pipeline phase and the most specific applicable subcategory.
Repository review: Check formatting, links, publication labels, and possible duplicates before merging a new entry.

Benchmark names alone are not used to infer a paper's scope. For example, a paper is labeled as multi-document question answering only when its title, abstract, task definition, or experimental description explicitly requires evidence from multiple documents.

Duplicate Handling and Publication Status

Records are deduplicated in the following order: DOI, arXiv identifier, and normalized title. When a peer-reviewed version and an arXiv preprint refer to the same work, they are represented by one entry. The peer-reviewed venue and year are used as the publication label, while an arXiv link may be retained when it provides the most accessible full text.

Publication status is reported explicitly:

peer-reviewed papers use the confirmed conference or journal name and year;
Findings, industry-track, workshop, and demo papers retain their specific track designation when applicable; and
papers without a confirmed peer-reviewed publication are labeled ArXiv rather than being assigned a predicted venue.

If a venue changes or a preprint is subsequently published, the existing record is updated instead of adding a second entry.

Classification Procedure

The taxonomy follows the stage of the LLM pipeline in which the graph makes its primary contribution:

Input phase: graphs construct prompts, supply external knowledge, or organize retrieval. We distinguish a pre-existing knowledge graph, a reusable graph induced from a document corpus, and a temporary graph constructed for a particular input instance.
Model phase: graphs modify the processing of a single model or organize interactions among multiple models. Single-model methods are divided into external graph adapters and internal model fusion, while multi-model systems are divided into static and dynamic communication topologies.
Output phase: graphs organize or assess generated results. Reasoning graphs structure intermediate inference, planning graphs organize executable actions and dependencies, and evaluation graphs assess factuality or logicality.

The taxonomy is contribution-oriented rather than mutually exclusive. A system may combine retrieval, model fusion, agent coordination, reasoning, and planning. In such cases, the repository places the paper according to the graph mechanism emphasized as its primary contribution and may discuss its secondary roles in the survey text.

For the input phase, classification depends on the graph actually consumed by the method, not on the historical origin of every fact. A pre-built and reusable knowledge graph is classified as prompting from a knowledge graph; a graph index constructed from the target text collection is classified as a corpus-level graph; and a query-conditioned, disposable graph is classified as an instance-level graph.

Quality Control and Updates

Before an entry is accepted, its bibliographic identity and publication status are checked against an authoritative source. The title in this list follows the published paper, and links should resolve to an official proceedings page, publisher page, OpenReview record, or arXiv abstract. Classification decisions are checked against the paper's stated graph construction, graph–LLM interaction, and principal contribution rather than inferred only from keywords or benchmark names.

Because this repository is continuously maintained, omissions, metadata corrections, venue updates, and classification questions can be submitted through GitHub issues or pull requests. A proposed entry should provide the paper title, authors, publication status, URL, graph construction or source, role of the graph in the LLM pipeline, and suggested taxonomy location. Changes remain auditable through the repository's version history.

Name		Name	Last commit message	Last commit date
figs		figs
.gitignore		.gitignore
README.md		README.md

Latest commit

History

Folders and files

Repository files navigation