Preference Reconstruction on Real-Life Creative Problem Solving
77.05Preference AccuracyDICAI
Evaluation results
| Method | Links | |
|---|---|---|
| DICAIJudge LLM=GPT-52026.06 | 77.05 | |
| ICAIJudge LLM=GPT-52026.06 | 64.2 | |
| AutoRubricJudge LLM=GPT-52026.06 | 43.65 | |
| CoTJudge LLM=GPT-52026.06 | 38.37 | |
| ToTJudge LLM=GPT-52026.06 | 35.18 | |
| CoT-SCJudge LLM=GPT-52026.06 | 34.33 | |
| Self-RefineJudge LLM=GPT-52026.06 | 32.11 |