Knowledge Base Question Answering On Grailqa
Metrics
Compositional EM
Compositional F1
I.I.D. EM
I.I.D. F1
Overall EM
Overall F1
Zero-shot EM
Zero-shot F1
Results
Performance results of various models on this benchmark
Model Name | Compositional EM | Compositional F1 | I.I.D. EM | I.I.D. F1 | Overall EM | Overall F1 | Zero-shot EM | Zero-shot F1 | Paper Title | Repository |
|---|---|---|---|---|---|---|---|---|---|---|
| ReTraCk | 61.5 | 70.9 | 84.4 | 87.5 | 58.1 | 65.3 | 44.6 | 52.5 | ReTraCk: A Flexible and Efficient Framework for Knowledge Base Question Answering | |
| PoG-GPT4 (Tan et al., 2024) | - | - | - | - | - | - | - | - | Paths-over-Graph: Knowledge Graph Empowered Large Language Model Reasoning |
0 of 2 row(s) selected.