Story Completion on XStoryCloze 1.0 (test)
71.5XStoryCloze Accuracy (en)TokAlign
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| TokAlignBackbone=Pythia 6.9B, Tuning Stage=1k steps tuning on multilingual corpus2026.05 | 71.5 | 54.5 | 48.4 | 54.2 | |
| TokAlign++Backbone=Pythia 6.9B, Tuning Stage=1k steps tuning on multilingual corpus2026.05 | 71.5 | 55.1 | 48.2 | 54.4 | |
| PythiaBackbone=Pythia 6.9B, Tuning Stage=Vanilla (Zero-shot)2026.05 | 70.4 | 53.9 | 50.3 | 53.8 | |
| ZeTTBackbone=Pythia 6.9B, Tuning Stage=1k steps tuning on multilingual corpus2026.05 | 70.2 | 50.4 | 47.7 | 53.9 | |
| FocusBackbone=Pythia 6.9B, Tuning Stage=1k steps tuning on multilingual corpus2026.05 | 66.8 | 48.7 | 47.4 | 53.9 | |
| TokAlign++Backbone=Pythia 6.9B, Tuning Stage=Initialization without any tuning2026.05 | 66.5 | 48.4 | 46.5 | 50.2 | |
| TokAlignBackbone=Pythia 6.9B, Tuning Stage=Initialization without any tuning2026.05 | 64.7 | 47.7 | 47.2 | 50.2 | |
| TokAlign++Backbone=Pythia 1B, Tuning Stage=1k steps tuning on multilingual corpus2026.05 | 64.6 | 51.4 | 48 | 53.5 | |
| PythiaBackbone=Pythia 1B, Tuning Stage=Vanilla (Zero-shot)2026.05 | 64.3 | 48.6 | 48 | 52.9 | |
| ZeTTBackbone=Pythia 6.9B, Tuning Stage=Initialization without any tuning2026.05 | 64.3 | 47.3 | 47.6 | 49.4 | |
| TokAlignBackbone=Pythia 1B, Tuning Stage=1k steps tuning on multilingual corpus2026.05 | 63.4 | 49.8 | 47.6 | 53.2 | |
| ZeTTBackbone=Pythia 1B, Tuning Stage=1k steps tuning on multilingual corpus2026.05 | 62.5 | 49.4 | 47.5 | 53.1 | |
| TokAlign++Backbone=Pythia 1B, Tuning Stage=Initialization without any tuning2026.05 | 61 | 48.1 | 47.9 | 50.6 | |
| FocusBackbone=Pythia 1B, Tuning Stage=1k steps tuning on multilingual corpus2026.05 | 61 | 48.3 | 47.1 | 54.2 | |
| TokAlignBackbone=Pythia 1B, Tuning Stage=Initialization without any tuning2026.05 | 59.8 | 47.5 | 47.4 | 50.2 | |
| ZeTTBackbone=Pythia 1B, Tuning Stage=Initialization without any tuning2026.05 | 55.8 | 47.2 | 47.3 | 49.4 | |
| FocusBackbone=Pythia 6.9B, Tuning Stage=Initialization without any tuning2026.05 | 45.5 | 46.7 | 45.9 | 50.1 | |
| FocusBackbone=Pythia 1B, Tuning Stage=Initialization without any tuning2026.05 | 44.7 | 47.8 | 46.4 | 49.4 |