Knowledge Base Question Answering on CWQ (ComplexWebQuestions)
82.6Hits@1 AccuracyGraSP
Evaluation Results
| Method | Links | |
|---|---|---|
| GraSP2026.04 | 82.6 | |
| LMP2026.04 | 82.2 | |
| Chain-of-QuestionMethodology=KG + LLM with Fine-Tuning (Small Models Fine-Tuning)2026.04 | 78.8 | |
| KG-ReasonerSize=30B, Methodology=KG + LLM with Fine-Tuning (LLM Backbone Fine-Tuning)2026.04 | 78.14 | |
| PoG with GPT-4Methodology=KG + LLMs w/o Fine-Tuning2026.04 | 75 | |
| iQUEST2026.04 | 73.8 | |
| KG-AgentMethodology=KG + LLM with Fine-Tuning (LLM Backbone Fine-Tuning)2026.04 | 72.2 | |
| KG-Agent2026.04 | 72.2 | |
| KBQA-o12026.04 | 72 | |
| ToG2026.04 | 69.5 | |
| ToG with GPT 4Methodology=KG + LLMs w/o Fine-Tuning2026.04 | 67.6 | |
| Readi with GPT-4Methodology=KG + LLMs w/o Fine-Tuning2026.04 | 67 | |
| KBQA-o1Methodology=KG + LLM with Fine-Tuning (LLM Backbone Fine-Tuning)2026.04 | 67 | |
| RoGSize=7B, Methodology=KG + LLM with Fine-Tuning (LLM Backbone Fine-Tuning)2026.04 | 62.6 | |
| KG-CoT with GPT 4Methodology=KG + LLM with Fine-Tuning (Small Models Fine-Tuning)2026.04 | 62.3 | |
| LightPROF (LLaMA3-8B)Size=8B, Methodology=KG + LLM with Fine-Tuning (Small Models Fine-Tuning)2026.04 | 59.3 | |
| GPT-4o-mini + KGMethodology=KG + LLMs w/o Fine-Tuning2026.04 | 54.35 | |
| DeepSeek-R1-Distill-Llama-70B + KGSize=70B, Methodology=KG + LLMs w/o Fine-Tuning2026.04 | 52.38 | |
| Qwen-2.5-7B (SFT) + KGSize=7B, Methodology=KG + LLM with Fine-Tuning (LLM Backbone Fine-Tuning)2026.04 | 51.84 | |
| LLaMA-3.1-8B (SFT) + KGSize=8B, Methodology=KG + LLM with Fine-Tuning (LLM Backbone Fine-Tuning)2026.04 | 47.4 | |
| FlexKBQAMethodology=KG + LLM with Fine-Tuning (Small Models Fine-Tuning)2026.04 | 46.2 | |
| LLaMA-3.1-8B + KGSize=8B, Methodology=KG + LLMs w/o Fine-Tuning2026.04 | 45.64 | |
| Qwen-2.5-7B + KGSize=7B, Methodology=KG + LLMs w/o Fine-Tuning2026.04 | 44.82 | |
| LLaMA-3.3-70B + KGSize=70B, Methodology=KG + LLMs w/o Fine-Tuning2026.04 | 44 | |
| Interactive-KBQA (13B)Size=13B, Methodology=KG + LLM with Fine-Tuning (LLM Backbone Fine-Tuning)2026.04 | 42.5 | |
| GPT-4o-miniMethodology=LLM w/o KG2026.04 | 42.32 | |
| GPT-4oMethodology=LLM w/o KG2026.04 | 41.77 | |
| Interactive-KBQA (7B)Size=7B, Methodology=KG + LLM with Fine-Tuning (LLM Backbone Fine-Tuning)2026.04 | 39.9 | |
| LLaMA-3.3-70BSize=70B, Methodology=LLM w/o KG2026.04 | 37.2 | |
| LLaMA-3.1-8BSize=8B, Methodology=LLM w/o KG2026.04 | 32.33 | |
| DeepSeek-R1-Distill-Llama-70BSize=70B, Methodology=LLM w/o KG2026.04 | 31.92 | |
| Qwen-2.5-7BSize=7B, Methodology=LLM w/o KG2026.04 | 31.25 |