Pairwise LLM Evaluation on Hotels
82Win Rate (Contrast)Q-STRUM Debate
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Q-STRUM DebateComparison Baseline=Q-STRUM Base, Judge Model=Deepseek-v32025.02 | 82 | 55 | 83 | 80 | |
| Q-STRUM DebateComparison Baseline=Q-STRUM Contrastive, Judge Model=Deepseek-v32025.02 | 78 | 55 | 76 | 75 |