Knowledge Graph Question Answering on WebQSP 3-hop
0.819AccuracyCa2KG
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Ca2KGBase Model=GPT-3.52026.01 | 0.819 | 0.108 | 0.128 | 0.92 | |
| RC-RAGBase Model=GPT-3.52026.01 | 0.794 | 0.139 | 0.144 | 0.863 | |
| IoEBase Model=GPT-3.52026.01 | 0.673 | — | — | — | |
| Verb2S-CoTBase Model=GPT-3.52026.01 | 0.67 | 0.539 | 0.434 | 0.486 | |
| Ca2KGBase Model=LLaMA-32026.01 | 0.612 | 0.242 | 0.266 | 0.657 | |
| IoEBase Model=LLaMA-32026.01 | 0.606 | — | — | — | |
| Verb2S-CoTBase Model=LLaMA-32026.01 | 0.501 | 0.619 | 0.602 | 0.392 | |
| Verb1S-Top4Base Model=GPT-3.52026.01 | 0.492 | 0.449 | 0.448 | 0.437 | |
| Verb1S-Top4Base Model=LLaMA-32026.01 | 0.397 | 0.486 | 0.481 | 0.445 | |
| RC-RAGBase Model=LLaMA-32026.01 | 0.315 | 0.354 | 0.368 | 0.392 | |
| Verb2S-Top4Base Model=LLaMA-32026.01 | 0.315 | 0.354 | 0.368 | 0.392 | |
| Verb2S-Top4Base Model=GPT-3.52026.01 | 0.176 | 0.676 | 0.586 | 0.155 | |
| Self-CorrectBase Model=LLaMA-32026.01 | 0.165 | — | — | — | |
| Self-CorrectBase Model=GPT-3.52026.01 | 0.162 | — | — | — | |
| KG-RAGBase Model=GPT-3.52026.01 | 0.031 | 0.908 | 0.911 | 0.05 | |
| KG-RAGBase Model=LLaMA-32026.01 | 0.004 | 0.739 | 0.551 | 0.004 |