Preference Reconstruction on Experiment Design
81.77Preference AccuracyDICAI
Evaluation results
| Method | Links | |
|---|---|---|
| DICAIJudge LLM=GPT-52026.06 | 81.77 | |
| DICAIJudge Model=GPT-4o2026.06 | 79.93 | |
| ICAIJudge Model=GPT-4o2026.06 | 73.8 | |
| ICAIJudge LLM=GPT-52026.06 | 73.4 | |
| CoTJudge LLM=GPT-52026.06 | 66.22 | |
| CoTJudge Model=GPT-4o2026.06 | 65.66 | |
| CoT-SCJudge LLM=GPT-52026.06 | 64.96 | |
| ToTJudge LLM=GPT-52026.06 | 64.87 | |
| Self-RefineJudge Model=GPT-4o2026.06 | 64.53 | |
| CoT-SCJudge Model=GPT-4o2026.06 | 63.31 | |
| Self-RefineJudge LLM=GPT-52026.06 | 62.52 | |
| ToTJudge Model=GPT-4o2026.06 | 62.41 | |
| AutoRubricJudge LLM=GPT-52026.06 | 57.12 | |
| AutoRubricJudge Model=GPT-4o2026.06 | 55.46 |