Automated Essay Scoring on LAILA (cross-prompt)
55.7Relevance ScoreTRATES
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| TRATESInference Time=30 seconds per essay2026.03 | 55.7 | 69.6 | 65.7 | 66.4 | 65.2 | 60.8 | 64.3 | 74.4 | 65.3 | |
| MOOSEInference Time=1 second per essay2026.03 | 41.1 | 62.7 | 64.2 | 64.9 | 58.5 | 58.6 | 62.3 | 64.9 | 59.7 | |
| XGBModel Type=Feature-based, Inference Time=0.2 seconds per essay2026.03 | 36 | 64.5 | 64.1 | 64.1 | 58.3 | 57.7 | 61.9 | 67.9 | 59.3 | |
| NNModel Type=Feature-based, Inference Time=0.2 seconds per essay2026.03 | 35.3 | 60.9 | 62.1 | 63.1 | 56.6 | 56.5 | 59.7 | 65.1 | 57.4 | |
| RFModel Type=Feature-based, Inference Time=0.3 seconds per essay2026.03 | 33.1 | 60.9 | 64.4 | 63.7 | 57.3 | 55.9 | 60.9 | 68.2 | 58.1 |