Multi-Constraint Text Generation on CompMCTG Average 1.0
4.02Relevance (avg)Con. Prefix
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Con. PrefixMethod Paradigm=Joint-Training based2024.04 | 4.02 | 3.26 | |
| DCGMethod Paradigm=Joint-Training based2024.04 | 3.79 | 3.74 | |
| CTRLMethod Paradigm=Joint-Training based2024.04 | 3.77 | 3.88 | |
| Dis-LensMethod Paradigm=Separate-Training based2024.04 | 3.55 | 3.02 | |
| PriorMethod Paradigm=Separate-Training based2024.04 | 3.16 | 3.09 | |
| CatPromptMethod Paradigm=Joint-Training based2024.04 | 3 | 3.45 | |
| LLaMA-2Method Paradigm=LLM+In-Context Learning2024.04 | 2.85 | 4.53 | |
| ChatGPTMethod Paradigm=LLM+In-Context Learning2024.04 | 2.71 | 4.79 | |
| FudgeMethod Paradigm=Decoding-Time based2024.04 | 2.44 | 2.1 | |
| PPLMMethod Paradigm=Decoding-Time based2024.04 | 2.06 | 1.09 |