Multi-aspect Controllable Text Generation on CompMCTG 1.0 (Hold-Out)
84.84Aid ScoreDis-Lens
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Dis-LensApproach Type=Separate-Training based2024.04 | 84.84 | 95.84 | 55.58 | 104.89 | — | — | — | — | |
| Con. PrefixApproach Type=Joint-Training based2024.04 | 83.75 | 80.49 | 80.36 | 83.9 | — | — | — | — | |
| DCGApproach Type=Joint-Training based2024.04 | 79.72 | 62.05 | 76.66 | 64.4 | — | — | — | — | |
| CTRLApproach Type=Joint-Training based2024.04 | 78.89 | 51.2 | 75.09 | 51.28 | — | — | — | — | |
| PriorApproach Type=Separate-Training based2024.04 | 73.64 | 108.58 | 49.93 | 97.64 | — | — | — | — | |
| CatPromptApproach Type=Joint-Training based2024.04 | 63.95 | 73.24 | 60.32 | 68.45 | — | — | — | — | |
| LLAMA-2Approach Type=LLM+In-context Learning2024.04 | 62.61 | 25.55 | 40.82 | 23.8 | — | — | — | — | |
| FudgeApproach Type=Decoding-Time based2024.04 | 59.35 | 179.47 | 42.1 | 252.08 | — | — | — | — | |
| ChatGPTApproach Type=LLM+In-context Learning2024.04 | 56.62 | 18.29 | 49.21 | 18.49 | — | — | — | — | |
| PPLMApproach Type=Decoding-Time based2024.04 | 41.05 | 325.09 | 40.62 | 340.76 | — | — | — | — | |
| CatPromptMethod Paradigm=Joint-Training based2024.04 | — | — | — | — | 3.27 | 3.49 | 3.04 | 3.58 | |
| ChatGPTMethod Paradigm=LLM+In-Context Learning2024.04 | — | — | — | — | 2.86 | 4.75 | 2.47 | 4.81 | |
| Con. PrefixMethod Paradigm=Joint-Training based2024.04 | — | — | — | — | 4.19 | 3.4 | 4.01 | 3.13 | |
| CTRLMethod Paradigm=Joint-Training based2024.04 | — | — | — | — | 3.78 | 3.92 | 3.75 | 3.94 | |
| DCGMethod Paradigm=Joint-Training based2024.04 | — | — | — | — | 3.9 | 3.68 | 3.84 | 3.64 | |
| Dis-LensMethod Paradigm=Separate-Training based2024.04 | — | — | — | — | 4.1 | 3.12 | 2.55 | 3.01 | |
| FudgeMethod Paradigm=Decoding-Time based2024.04 | — | — | — | — | 2.68 | 2.13 | 2.07 | 1.87 | |
| LLaMA-2Method Paradigm=LLM+In-Context Learning2024.04 | — | — | — | — | 3.23 | 4.48 | 2.37 | 4.43 | |
| PPLMMethod Paradigm=Decoding-Time based2024.04 | — | — | — | — | 2.22 | 1.07 | 2.01 | 1.09 | |
| PriorMethod Paradigm=Separate-Training based2024.04 | — | — | — | — | 3.53 | 3.04 | 2.42 | 3.2 |