Multitask NLP on DecaNLP Subset (SQuAD 2.0, WikiSQL, SST, QA-SRL, WOZ) (test)
78.2Average ScoreMultitasked
Evaluation Results
| Method | Links | |
|---|---|---|
| Multitaskeddescription=Upper bound for comparison2025.12 | 78.2 | |
| PPSEBMsampling_rate=0.2, mode=GEN2025.12 | 77.4 | |
| LSEBMCLsampling_rate=0.2, mode=GEN2025.12 | 77.3 | |
| HMI-LAMOLsampling_rate=0.2, mode=GEN2025.12 | 76.9 | |
| PPSEBMsampling_rate=0.05, mode=GEN2025.12 | 76.7 | |
| LSEBMCLsampling_rate=0.05, mode=GEN2025.12 | 76.5 | |
| HMI-LAMOLsampling_rate=0.05, mode=GEN2025.12 | 76 | |
| LAMOLsampling_rate=0.2, mode=GEN2025.12 | 74.1 | |
| LAMOLsampling_rate=0.05, mode=GEN2025.12 | 70.3 | |
| Fine-tuned2025.12 | 52.6 | |
| MAS2025.12 | 51.2 |