Critique Quality Evaluation on Summarization
75Win RateRCO
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| RCOComparison Target=LLaMa-3-70B-Instruct2025.06 | 75 | 15 | 10 | 65 | |
| RCOComparison Target=LLaMa-3-8B-Instruct2025.06 | 60 | 25 | 15 | 45 | |
| RCOComparison Target=LLaMa-3-8B-Instruct2025.06 | 57.5 | 17.5 | 25 | 32.5 | |
| RCOComparison Target=LLaMa-3-8B-Instruct + MSE + DPO2025.06 | 52.5 | 15 | 32.5 | 20 | |
| RCOComparison Target=LLaMa-3-70B-Instruct2025.06 | 40 | 45 | 15 | 25 | |
| RCOComparison Target=LLaMa-3-8B-Instruct + MSE + DPO2025.06 | 30 | 50 | 20 | 10 |