CTT-difficulty prediction on BEA (466/201 split)
0.309Standard Deviation (SD)Shared-task winner
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Shared-task winnerntrain / ntest=466 / 201, design=leaderboard2026.07 | 0.309 | 0.299 | 0.97 | — | |
| Dummy (mean) baselinentrain / ntest=466 / 201, design=leaderboard2026.07 | 0.309 | 0.311 | 1.01 | 0 | |
| Fine-tuned ensemblentrain / ntest=466 / 201, design=leaderboard2026.07 | 0.309 | 0.2926 | 0.95 | — |