Knowledge Graph Completion on NYC-Large (test)
52GPT AccuracyUrbanKGent-13B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| UrbanKGent-13BModel Scale=13B2024.02 | 52 | 3.67 | 46 | |
| UrbanKGent InferenceMode=Inference setting2024.02 | 51 | 3.87 | 43 | |
| UrbanKGent-8BModel Scale=8B2024.02 | 51 | 4.16 | 44 | |
| UrbanKGent-7BModel Scale=7B2024.02 | 50 | 4.07 | 44 | |
| ICLLearning Paradigm=In-context-learning2024.02 | 48 | 4.15 | 39 | |
| VFTLearning Paradigm=Vanilla Fine-tuning2024.02 | 46 | 3.98 | 35 | |
| ZSLLearning Paradigm=Zero-shot-learning2024.02 | 45 | 3.95 | 31 |