Preference Reconstruction on Hypothesis Generation
72.8Preference AccuracyDICAI
Evaluation Results
| Method | Links | |
|---|---|---|
| DICAIJudge LLM=GPT-52026.06 | 72.8 | |
| ICAIJudge LLM=GPT-52026.06 | 66.4 | |
| AutoRubricJudge LLM=GPT-52026.06 | 56.79 | |
| CoTJudge LLM=GPT-52026.06 | 56.1 | |
| CoT-SCJudge LLM=GPT-52026.06 | 55.16 | |
| ToTJudge LLM=GPT-52026.06 | 54.13 | |
| Self-RefineJudge LLM=GPT-52026.06 | 48.43 |