Node Classification on Cora (Accuracy and Macro-F1)
79.31Macro-F1LoReC
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LoReCBase Model=GraphPrompter, Backbone=Llama2-7b2026.04 | 79.31 | 81.37 | |
| LoReCBase Model=GraphPrompter, Backbone=Qwen3-8b2026.04 | 77.62 | 80.26 | |
| GraphPrompterBackbone=Qwen3-8b2026.04 | 77.27 | 80.26 | |
| UniGraphLM2026.05 | 75.72 | 76.98 | |
| LLaGA2026.05 | 73.29 | 74.71 | |
| GPT-4oType=LLM, Zero-shot=true2026.04 | 71.44 | 70.3 | |
| GraphPrompterBackbone=Llama2-7b2026.04 | 71.34 | 79.15 | |
| TRN-R1-Zero (7B)Type=RL Only, Zero-shot=true2026.04 | 70.33 | 72.59 | |
| Qwen2.5-14B-itType=LLM, Zero-shot=true2026.04 | 68.26 | 67.22 | |
| Graph-R1 (14B)Type=SFT + RL, Zero-shot=true2026.04 | 67.34 | 68.15 | |
| Qwen2.5-7B-itType=LLM, Zero-shot=true2026.04 | 67.19 | 67.59 | |
| Llama-3.1-8BType=LLM, Zero-shot=true2026.04 | 64.41 | 64.55 | |
| GOFA2026.05 | 60.88 | 63.88 | |
| TEA-GLM2026.05 | 59.71 | 62.38 | |
| Vicuna-7B2026.05 | 58.52 | 56.67 | |
| GraphGPT2026.05 | 57.91 | 57.59 | |
| ZeroGType=GFM, Zero-shot=true2026.04 | 57.56 | 62.55 | |
| Qwen3-8bBackbone=Qwen3-8b2026.04 | 56.47 | 65.68 | |
| Qwen2.5-1.5B-itType=LLM, Zero-shot=true2026.04 | 49.91 | 47.96 | |
| CE-FedGNN2026.05 | 47.01 | — | |
| FedGNN-ST2026.05 | 43.45 | — | |
| Swift2026.05 | 42.96 | — | |
| FedGCN2026.05 | 40.05 | — | |
| LLaMA-2-7B2026.05 | 30.13 | 29.45 | |
| Llama2-7bBackbone=Llama2-7b2026.04 | 30.02 | 37.64 | |
| FedPUB2026.05 | 27.82 | — | |
| SC2026.05 | 25.11 | — | |
| FedSage+2026.05 | 18.76 | — | |
| FedAvg2026.05 | 18.68 | — | |
| LLaGAType=GFM, Zero-shot=true2026.04 | 8.49 | 18.82 |