Scientific Modeling Idea Generation on Scientific Modeling Ideas 12 evaluation instances
66Wilcoxon StatisticDN-Hypo-Pipeline
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DN-Hypo-PipelineEvaluator=LLM-as-Judge, Null Hypothesis=H0(1): Max score comparison2026.06 | 66 | 0.0005 | |
| DN-Hypo-PipelineEvaluator=LLM-as-Judge, Null Hypothesis=H0(2): Mean score comparison2026.06 | 63.5 | 0.0271 | |
| DN-Hypo-PipelineEvaluator=LLM-as-Judge, Null Hypothesis=H0(3): Median score comparison2026.06 | 28.5 | 0.0859 |