Multi-hop Question Answering on Frames
41.38ACCETC+FM*
Evaluation results
| Method | Links | ||
|---|---|---|---|
| TC+FM*Model=Qwen-72B2026.02 | 41.38 | 58.59 | |
| Search-o1Model=Qwen-72B2026.02 | 37.37 | 56.67 | |
| TC+FM*Model=Qwen-32B2026.02 | 37.15 | 53.86 | |
| TC+FM*Model=Qwen-14B2026.02 | 34.83 | 49.51 | |
| Search-o1Model=Qwen-32B2026.02 | 34.34 | 54.12 | |
| Search-o1Model=Qwen-14B2026.02 | 30.46 | 46.48 | |
| TC+FM*Model=Llama3.1-8B2026.02 | 27.43 | 39.92 | |
| Standard RAGModel=Qwen-72B2026.02 | 25.97 | 37.62 | |
| TC+FM*Model=Qwen-7B2026.02 | 25.12 | 34.71 | |
| Search-o1Model=Llama3.1-8B2026.02 | 24.88 | 37.62 | |
| Search-o1Model=Qwen-7B2026.02 | 24.63 | 38.34 | |
| Naive GenModel=Qwen-72B2026.02 | 21.12 | 32.16 | |
| Naive GenModel=Llama3.1-8B2026.02 | 21.12 | 32.16 | |
| Standard RAGModel=Qwen-32B2026.02 | 19.78 | 30.95 | |
| TC+FM*Model=Qwen-3B2026.02 | 18.84 | 19.42 | |
| Standard RAGModel=Qwen-14B2026.02 | 17.96 | 28.4 | |
| Standard RAGModel=Qwen-7B2026.02 | 15.78 | 24.76 | |
| Naive GenModel=Qwen-32B2026.02 | 15.66 | 24.51 | |
| Standard RAGModel=Llama3.1-8B2026.02 | 15.05 | 22.82 | |
| Naive GenModel=Qwen-14B2026.02 | 14.81 | 22.82 | |
| Standard RAGModel=Qwen-3B2026.02 | 12.01 | 19.17 | |
| Search-o1Model=Qwen-3B2026.02 | 11.77 | 18.81 | |
| Naive GenModel=Qwen-7B2026.02 | 11.29 | 16.87 | |
| Naive GenModel=Qwen-3B2026.02 | 5.94 | 8.98 |