Automated Essay Scoring on ASAP Prompt 1 All essays (cross-validation)
0.856Quadratic Weighted Kappa (QWK)NPCR
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| NPCR2026.04 | 0.856 | — | — | — | — | |
| BERTarchitecture=Standard BERT, validation=5-fold CV2026.04 | 0.803 | — | — | — | 0.809 | |
| Human-human QWK2026.04 | — | — | — | 0.721 | — | |
| Human-like ceiling2026.04 | — | — | 0.777 | — | — | |
| Theoretical ceiling2026.04 | — | 0.915 | — | — | — |