Topic Writing on LaMP Topic Writing
0.291ROUGE-1C-BPO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| C-BPOTraining Paradigm=Preference Opt.2026.05 | 0.291 | 0.118 | |
| CoPETraining Paradigm=Preference Opt.2026.05 | 0.281 | 0.12 | |
| OPPUTraining Paradigm=SFT-based2026.05 | 0.278 | 0.112 | |
| BCOTraining Paradigm=Preference Opt.2026.05 | 0.272 | 0.112 | |
| KTOTraining Paradigm=Preference Opt.2026.05 | 0.269 | 0.109 | |
| PAGTraining Paradigm=Non-Tuned2026.05 | 0.262 | 0.107 | |
| TAMTraining Paradigm=SFT-based2026.05 | 0.253 | 0.107 | |
| Base ModelTraining Paradigm=Non-Tuned2026.05 | 0.246 | 0.105 | |
| RAGTraining Paradigm=Non-Tuned2026.05 | 0.243 | 0.115 |