List your company

Node Classification on PubMed 10-shot

92.77AccuracyLG Co-Teaching
66.270873.150480.0386.9096Dec 9, 2025Jan 8, 2026Feb 8, 2026Mar 10, 2026Apr 10, 2026May 10, 2026Jun 10, 2026

Evaluation results

MethodLinks
LG Co-Teachingshot setting=10-shot2026.0692.77—
GNN-as-Judgeshot setting=10-shot2026.0690.17—
HopRankrequires_labeled_nodes_during_training=false2026.0488.62—
Graph-CoTshot setting=10-shot2026.0686.9—
InstructTuningrequires_labeled_nodes_during_training=true, Backbone=LLaMA-3-8B2026.0485.03—
w/ Few-Shotrequires_labeled_nodes_during_training=true, Backbone=LLaMA-3-8B2026.0479.75—
TAPEshot setting=10-shot2026.0677.18—
LLM-GNNshot setting=10-shot2026.0676.82—
GCNshot setting=10-shot2026.0676.24—
GCNrequires_labeled_nodes_during_training=true2026.0475.33—
Graph-CoT w. Neighborshot setting=10-shot2026.0675.26—
GraphSAGErequires_labeled_nodes_during_training=true2026.0475.12—
SGCrequires_labeled_nodes_during_training=true2026.0474.98—
w/ Neighborsrequires_labeled_nodes_during_training=false, Backbone=LLaMA-3-8B2026.0474.98—
SAGEshot setting=10-shot2026.0674.88—
GATshot setting=10-shot2026.0674.87—
Zero-Shotshot setting=10-shot2026.0674.78—
Zero-Shotrequires_labeled_nodes_during_training=false, Backbone=LLaMA-3-8B2026.0474.51—
GLEMshot setting=10-shot2026.0674.17—
GATrequires_labeled_nodes_during_training=true2026.0472.84—
GPF-plusPre-training strategy=MaskedEdge2025.1272.0866.18
LEAPPre-training strategy=ContraEdge2025.1271.8466.95
LEAPPre-training strategy=MaskedEdge2025.1271.6866.65
RELIEFPre-training strategy=ContraEdge2025.1271.3566.82
GraphGPTrequires_labeled_nodes_during_training=true, Backbone=LLaMA-3-8B2026.0471.33—
GraphGPTshot setting=10-shot2026.0671.33—
SUPTsoftPre-training strategy=MaskedEdge2025.1271.1865.9
GPF-plusPre-training strategy=ContraEdge2025.1271.1866.39
RELIEFPre-training strategy=MaskedEdge2025.1271.1666.37
SUPTsoftPre-training strategy=ContraEdge2025.1270.9966.08
FTPre-training strategy=MaskedEdge2025.1270.9165.42
GPFPre-training strategy=ContraEdge2025.1270.9165.99
SUPThardPre-training strategy=ContraEdge2025.1270.8666.23
SUPThardPre-training strategy=MaskedEdge2025.1270.5965.29
FTPre-training strategy=ContraEdge2025.1270.1565.84
GPPTPre-training strategy=MaskedEdge2025.1270.1165.05
GPFPre-training strategy=MaskedEdge2025.1269.8164.97
GPromptPre-training strategy=ContraEdge2025.1269.564.73
LLaGArequires_labeled_nodes_during_training=true, Backbone=LLaMA-3-8B2026.0467.29—
LLaGAshot setting=10-shot2026.0667.29—