Value Alignment on Moral Stories (test)
4.85Align ScoreAEM + VM
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| AEM + VMFramework=SECOND THOUGHTS2023.01 | 4.85 | 26.73 | 5.41 | 11.96 | |
| AEM + AILFramework=SECOND THOUGHTS2023.01 | 4.55 | 25.2 | 5.64 | 9.23 | |
| DExperts2023.01 | 4.47 | 22.41 | 5.35 | 6.28 | |
| InstructGPTSource=Huge LM API service2023.01 | 4.2 | 25.4 | 5.69 | 5.38 | |
| AEM OnlyFramework=SECOND THOUGHTS2023.01 | 3.8 | 24.1 | 5.22 | 10.55 | |
| GPT-3Source=Huge LM API service2023.01 | 3.28 | 22.26 | 5.34 | 7.31 | |
| Safe Beam Search2023.01 | 3.08 | 18.48 | 4.02 | 19.5 | |
| Data Filtering2023.01 | 2.7 | 13.32 | 4.43 | 7.94 | |
| MLE2023.01 | 2.48 | 7.96 | 4.54 | 8.26 | |
| PPLM2023.01 | 2.29 | 11.9 | 5.05 | 14.47 |