Node Classification on Products (Accuracy, Macro-F1)
77.67AccuracyLoReC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LoReCBase Model=GraphPrompter, Backbone=Llama2-7b2026.04 | 77.67 | 38.54 | |
| GraphPrompterBackbone=Llama2-7b2026.04 | 77.2 | 38.33 | |
| LoReCBase Model=GraphPrompter, Backbone=Qwen3-8b2026.04 | 74.59 | 37.72 | |
| GraphPrompterBackbone=Qwen3-8b2026.04 | 74.22 | 36.75 | |
| Qwen3-8bBackbone=Qwen3-8b2026.04 | 17.61 | 14.61 | |
| Llama2-7bBackbone=Llama2-7b2026.04 | 11.17 | 4.74 | |
| NOPE*coarsening ratio=0.5, category=Ours2026.05 | 0.679 | 0.6827 | |
| NOPEcoarsening ratio=0.5, category=Ours2026.05 | 0.6783 | 0.682 | |
| A-CMcoarsening ratio=0.5, category=Graph Coarsening2026.05 | 0.6755 | 0.6754 | |
| Degreecoarsening ratio=0.5, category=Full Graph2026.05 | 0.5859 | 0.6078 | |
| RAGcoarsening ratio=0.5, category=Full Graph2026.05 | 0.5789 | 0.6003 | |
| Randomcoarsening ratio=0.5, category=Full Graph2026.05 | 0.5761 | 0.598 | |
| UGCcoarsening ratio=0.5, category=Graph Coarsening2026.05 | 0.5484 | 0.5676 |