| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Commonsense Reasoning | SocialIQA | Accuracy88.1 | 97 | |
| Social Commonsense Reasoning | SocialIQA | Accuracy87.11 | 68 | |
| Commonsense Question Answering | SocialIQA (SIQA) (val) | Accuracy70.7 | 24 | |
| Question Answering | SocialIQA | Accuracy83.9 | 16 | |
| Ranking correlation with full dataset evaluation | SocialIQA | Kendall Correlation0.81 | 10 | |
| Scaling Law Prediction | SocialIQA | MAE0.0088 | 7 | |
| Preference alignment | SocialIQA | Preference Alignment87.3 | 5 | |
| Adaptivity | SocialIQA | Adaptivity75 | 4 | |
| Commonsense Reasoning | SOCIALIQA (dev) | Accuracy73.8 | 3 |