List your company

Multi-hop Question Quality Evaluation on Bridge Questions

98.9Multi-Hop AccuracyHopWeaver
89.43691.89394.3596.807May 21, 2025

Evaluation results

MethodLinks
HopWeaverLLM Engine=QwQ-32B2025.0598.94.23
HopWeaverLLM Engine=Qwen3-14B2025.0596.94.09
HopWeaverLLM Engine=Gemini-2.5-flash2025.0596.44.27
HotpotQAGeneration Source=Human Datasets (Baselines)2025.0592.84.23
2WikiMultiHopQAGeneration Source=Human Datasets (Baselines)2025.0592.84.04
MuSiQueGeneration Source=Human Datasets (Baselines)2025.0591.23.78
HopWeaverLLM Engine=GLM-4-9B-04142025.0589.83.87