Summarization on Summarization benchmark dataset
85.28Win RateIPO*
Evaluation Results
| Method | Links | |
|---|---|---|
| IPO*Temperature=0.1, Evaluator=Qwen3-1.7B2026.02 | 85.28 | |
| IPO*Temperature=1.0, Evaluator=Qwen3-1.7B2026.02 | 83.87 | |
| DIPO+Temperature=0.1, Evaluator=Qwen3-1.7B2026.02 | 77.55 | |
| Sampled IPOTemperature=0.1, Evaluator=Qwen3-1.7B2026.02 | 73.07 | |
| DIPOTemperature=0.1, Evaluator=Qwen3-1.7B2026.02 | 71.71 | |
| DIPO+Temperature=1.0, Evaluator=Qwen3-1.7B2026.02 | 68.13 | |
| IPOTemperature=0.1, Evaluator=Qwen3-1.7B2026.02 | 66.19 | |
| Sampled IPOTemperature=1.0, Evaluator=Qwen3-1.7B2026.02 | 61.51 | |
| DIPOTemperature=1.0, Evaluator=Qwen3-1.7B2026.02 | 60.03 | |
| IPOTemperature=1.0, Evaluator=Qwen3-1.7B2026.02 | 54.71 | |
| SFTTemperature=0.1, Evaluator=Qwen3-1.7B2026.02 | 53.47 | |
| SFTTemperature=1.0, Evaluator=Qwen3-1.7B2026.02 | 39 |