Reproduce “GraphInfer-Bench: Benchmarking LLM's Inference Capability on Graphs” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhuoyi Peng, Jingzhou Jiang, Hanlin Gu, Lixin Fan, Yi Yang
GraphInfer-Bench evaluates whether models infer neighborhood-level properties from text-attributed graphs rather than retrieve facts from individual nodes or follow paths. It defines five description and comparison tasks across six domains, using sampled title-based ego-graphs. Results vary by task: GraphSAGE reaches 0.457 NMI on community partitioning, while Claude Opus 4.7 leads on outlier detection; results do not establish whole-graph reasoning or deployment.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.