Multi-hop Question Quality Evaluation on Bridge Questions
98.9Multi-Hop AccuracyHopWeaver
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HopWeaverLLM Engine=QwQ-32B2025.05 | 98.9 | 4.23 | |
| HopWeaverLLM Engine=Qwen3-14B2025.05 | 96.9 | 4.09 | |
| HopWeaverLLM Engine=Gemini-2.5-flash2025.05 | 96.4 | 4.27 | |
| HotpotQAGeneration Source=Human Datasets (Baselines)2025.05 | 92.8 | 4.23 | |
| 2WikiMultiHopQAGeneration Source=Human Datasets (Baselines)2025.05 | 92.8 | 4.04 | |
| MuSiQueGeneration Source=Human Datasets (Baselines)2025.05 | 91.2 | 3.78 | |
| HopWeaverLLM Engine=GLM-4-9B-04142025.05 | 89.8 | 3.87 |