Natural Language Understanding on GLUE (Custom Metrics)
80.9Red ScoreRMT-KD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RMT-KDBase Model=BERT-base2026.01 | 80.9 | 1.8 | |
| RMT-KDBase Model=BERT-tiny2026.01 | 58.8 | 1.4 | |
| DistilBERTBase Model=BERT-tiny2026.01 | 54.8 | 0.4 | |
| TheseusBase Model=BERT-tiny2026.01 | 53 | 0.1 | |
| PKDBase Model=BERT-tiny2026.01 | 50.1 | -0.8 | |
| TheseusBase Model=BERT-base2026.01 | 48.3 | 0.6 | |
| DistilBERTBase Model=BERT-base2026.01 | 42.7 | 0.2 | |
| PKDBase Model=BERT-base2026.01 | 40.5 | -1 |