Knowledge Graph Question Answering on MetaQA 1-hop 1.0 (test)
97.5Accuracy (%)KELP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| KELPfew-shot=4-shot, context=Knowledge Enhanced Link Prediction context2024.06 | 97.5 | 0.26 | |
| KELPfew-shot=12-shot, context=Knowledge Enhanced Link Prediction context2024.06 | 97.1 | 0.26 | |
| KELPfew-shot=8-shot, context=Knowledge Enhanced Link Prediction context2024.06 | 97 | 0.26 | |
| LLME (KG-GPT)few-shot=12-shot, context=LLM-based evidence2024.06 | 96.3 | 0.82 | |
| LLME (KG-GPT)few-shot=8-shot, context=LLM-based evidence2024.06 | 95.8 | 0.82 | |
| LLME (KG-GPT)few-shot=4-shot, context=LLM-based evidence2024.06 | 94.7 | 0.82 | |
| GPT (gpt-3.5-turbo-0613)few-shot=12-shot, context=Question-only2024.06 | 63 | 4.59 | |
| GPT (gpt-3.5-turbo-0613)few-shot=8-shot, context=Question-only2024.06 | 61.5 | 4.59 | |
| GPT (gpt-3.5-turbo-0613)few-shot=4-shot, context=Question-only2024.06 | 54.4 | 4.59 |