Creative Writing on Creative Writing Alpaca-Eval 100 problems 2.0
93.81Length-Controlled Win RateICRL prompting
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ICRL promptingComparison baseline=Best-of-N2025.05 | 93.81 | 1.01 | |
| ICRL promptingComparison baseline=Self-Refine2025.05 | 86.32 | 3.03 | |
| ICRL promptingComparison baseline=Long CoT2025.05 | 78.36 | 1.99 | |
| ICRL promptingComparison baseline=Reflexion2025.05 | 59.48 | 3.47 |