LINKX Benchmarks
Penn94
Structure
1Graphs
41,554Nodes
1,362,229Edges
2Classes
5Node Feature Dim
0.47Homophily
Coverage
90Results
1Variants
15Papers
- WWW 2026: Frequency-Corrupt Based Graph Self-Supervised Learning Frequency-Corrupt Based Graph Self-Supervised Learning WWW 2026
- KDD 2025: Discrepancy-Aware Graph Mask Auto-Encoder Discrepancy-Aware Graph Mask Auto-Encoder KDD 2025
- ICML 2024: GrokFormer: Graph Fourier Kolmogorov-Arnold Transformers GrokFormer: Graph Fourier Kolmogorov-Arnold Transformers ICML 2024
- ICML 2024: Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs ICML 2024
- Differentiable Cluster Graph Neural Network Differentiable Cluster Graph Neural Network
- WWW 2024: Optimizing Polynomial Graph Filters: A Novel Adaptive Krylov Subspace Approach Optimizing Polynomial Graph Filters: A Novel Adaptive Krylov Subspace Approach WWW 2024
- Simplified PCNet with Robustness Simplified PCNet with Robustness
- AAAI 2023: PC-Conv: Unifying Homophily and Heterophily with Two-fold Filtering PC-Conv: Unifying Homophily and Heterophily with Two-fold Filtering AAAI 2023
- ICLR 2023: Mixture of Weak & Strong Experts on Graphs Mixture of Weak & Strong Experts on Graphs ICLR 2023
- Clarify Confused Nodes via Separated Learning Clarify Confused Nodes via Separated Learning
- ICML 2023: Graph Neural Networks with Learnable and Optimal Polynomial Bases Graph Neural Networks with Learnable and Optimal Polynomial Bases ICML 2023
- Uplifting Message Passing Neural Network with Graph Original Information Uplifting Message Passing Neural Network with Graph Original Information
- Decoupled Self-supervised Learning for Non-Homophilou Graphs Decoupled Self-supervised Learning for Non-Homophilou Graphs
- ICML 2022: Finding Global Homophily in Graph Neural Networks When Meeting Heterophily Finding Global Homophily in Graph Neural Networks When Meeting Heterophily ICML 2022
- NeurIPS 2022: Convolutional Neural Networks on Graphs with Chebyshev Approximation, Revisited Convolutional Neural Networks on Graphs with Chebyshev Approximation, Revisited NeurIPS 2022
Binary node classification Facebook100 Penn94 social network node classification (binary gender). Uses the released sparse node features and a low-homophily graph. Feature-only baselines are essential because structure can be weakly predictive. Evaluated by Accuracy.
Dataset details
Task
Facebook100 Penn94 social network node classification (binary gender). The primary catalog metric is Accuracy.
Data
Facebook100 Penn94 social network node classification (binary gender). The cataloged artifact reports 1 graphs, 2 target classes or tasks.
Features
Each release supplies a sparse node-feature matrix, graph edges, and node labels. Feature meaning varies by source; structure and features are deliberately evaluated without assuming neighboring labels agree.
Splits and evaluation
The catalog records these protocols or variants: Standard split. Evaluation uses Accuracy (higher is better). Exact masks or folds must come from the cited release.
Quirks and pitfalls
These are transductive single-graph tasks with fixed random splits. Label homophily is low and class balance varies; feature-only and structure-only baselines are essential, and similarly named OGB/source datasets may use different labels or splits.
Sources
SOTA Progression
new SOTA
matches SOTA at that time
other proposed result
other baseline result
Leaderboard
| # | Model | Accuracy ↑ | Paper |
|---|---|---|---|
| 1 | DC-GNN | 0.8669
±
0.0022 | |
| 2 | M2M-GNN | 0.8594
±
0.0040 | |
| 3 | GloGNN | 0.8557 | |
| 4 | INGNN | 0.8544
±
0.0060 | |
| 5 | AdaptKry | 0.8497
±
0.0036 | |
| 6 | FavardGNN | 0.8492
±
0.0041 | |
| 7 | ChebNetII | 0.8486
±
0.0033 | |
| 8 | OptBasisGNN | 0.8485
±
0.0039 | |
| 9 | PCNet | 0.8475
±
0.0051 | |
| 10 | SPCNet-D | 0.8475
±
0.0027 | Simplified PCNet with Robustness
no code |
| 11 | LINKX | 0.8471
±
0.0052 | |
| 12 | MoE-GIN | 0.8456
±
0.0031 | |
| 13 | GPRGNN | 0.8455
±
0.0054 | |
| 14 | Specformer | 0.8432
±
0.0032 | |
| 15 | GOAL | 0.8418
±
0.0030 | |
| 16 | Ord. GNN | 0.8365
±
0.0060 | |
| 17 | MixHop | 0.8363
±
0.0054 | |
| 18 | GrokFormer | 0.8359
±
0.0026 | |
| 19 | AdaGNN | 0.8355
±
0.0031 | |
| 20 | GPR-GNN | 0.8354
±
0.0032 | |
| 21 | MoE-H2GCN | 0.8339
±
0.0043 | |
| 22 | JacobiConv | 0.8335
±
0.0011 | |
| 23 | BernNet | 0.8326
±
0.0029 | |
| 24 | H2GCN | 0.8320
±
0.0055 | |
| 25 | MoE-GCN | 0.8319
±
0.0043 | |
| 26 | GloGNN | 0.8307
±
0.0135 | |
| 27 | GCNII | 0.8292
±
0.0059 | |
| 28 | SPCNet-L | 0.8269
±
0.0022 | Simplified PCNet with Robustness
no code |
| 29 | GIN | 0.8268
±
0.0032 | |
| 30 | ACM-GCN | 0.8268
±
0.0060 | |
| 31 | ChebNet | 0.8259
±
0.0031 | |
| 32 | BernNet | 0.8247
±
0.0021 | |
| 33 | GCN | 0.8247
±
0.0027 | |
| 34 | AERO-GNN | 0.8247
±
0.0070 | |
| 35 | H_A | 0.8225 | |
| 36 | SIGN | 0.8213
±
0.0028 | |
| 37 | GraphMoE-GCN | 0.8161
±
0.0027 | |
| 38 | H2GCN | 0.8154
±
0.0056 | |
| 39 | GAT | 0.8145
±
0.0055 | |
| 40 | SGC | 0.8144
±
0.0015 | |
| 41 | GPRGNN | 0.8138
±
0.0016 | |
| 42 | H_2GCN | 0.8131
±
0.0060 | Simplified PCNet with Robustness
no code |
| 43 | GPR-GNN | 0.8112
±
0.0063 | |
| 44 | ChebyNet | 0.8109
±
0.0033 | |
| 45 | LINK | 0.8079
±
0.0049 | |
| 46 | GIN-skip | 0.8032
±
0.0043 | |
| 47 | MoE-GIN-skip | 0.8032
±
0.0043 | |
| 48 | PolyFormer | 0.7927
±
0.0026 | |
| 49 | MoE-SAGE | 0.7907
±
0.0043 | |
| 50 | GAT | 0.7768
±
0.0237 |
SOTA performance
set a new SOTA at the time
matched the SOTA at the time