Knowledge Base Question Answering on QALD en Freebase 10 (test)
74.28Hits@1KG-Hopper w/Qwen-2.5-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| KG-Hopper w/Qwen-2.5-7BSize=7B, Strategy=KG-Augmented with RL Fine-Tuning2026.03 | 74.28 | |
| GPT-4o-mini + KGStrategy=KG-Augmented without Fine-Tuning2026.03 | 72.86 | |
| LLaMA-3.3-70B + KGSize=70B, Strategy=KG-Augmented without Fine-Tuning2026.03 | 71.78 | |
| DeepSeek-R1-Distill-Llama-70B + KGSize=70B, Strategy=KG-Augmented without Fine-Tuning2026.03 | 66.1 | |
| Qwen-2.5-7B (SFT) + KGSize=7B, Strategy=KG-Augmented with Supervised Fine-Tuning2026.03 | 65.18 | |
| LLaMA-3.1-8B (SFT) + KGSize=8B, Strategy=KG-Augmented with Supervised Fine-Tuning2026.03 | 59.63 | |
| Qwen-2.5-7B + KGSize=7B, Strategy=KG-Augmented without Fine-Tuning2026.03 | 56.54 | |
| GPT-4oStrategy=LLM Prompting Only2026.03 | 56.2 | |
| LLaMA-3.3-70BSize=70B, Strategy=LLM Prompting Only2026.03 | 56 | |
| LLaMA-3.1-8B + KGSize=8B, Strategy=KG-Augmented without Fine-Tuning2026.03 | 55.73 | |
| GPT-4o-miniStrategy=LLM Prompting Only2026.03 | 51.98 | |
| DeepSeek-R1-Distill-Llama-70BSize=70B, Strategy=LLM Prompting Only2026.03 | 43.1 | |
| Qwen-2.5-7BSize=7B, Strategy=LLM Prompting Only2026.03 | 41.88 | |
| LLaMA-3.1-8BSize=8B, Strategy=LLM Prompting Only2026.03 | 40.25 |