Goal Prediction on FrozenLake OOD
90Classification AccuracyT-CoT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| T-CoTSize=4, Backbone=GPT-4o2026.06 | 90 | 56.7 | |
| SVoToSize=4, Backbone=Anole2026.06 | 88.3 | 63.3 | |
| SVoTpSize=4, Backbone=Anole2026.06 | 88.3 | 68.3 | |
| T-CoTSize=5, Backbone=GPT-4o2026.06 | 86.7 | 56.7 | |
| SVoTpSize=5, Backbone=Anole2026.06 | 85 | 58.3 | |
| SVoToSize=5, Backbone=Anole2026.06 | 83.3 | 56.7 | |
| MVoTSize=4, Backbone=Anole2026.06 | 80 | 56.7 | |
| T-CoTSize=6, Backbone=GPT-4o2026.06 | 76.7 | 53.3 | |
| SVoToSize=6, Backbone=Anole2026.06 | 76.7 | 40 | |
| SVoTpSize=6, Backbone=Anole2026.06 | 76.7 | 55 | |
| T-CoTSize=4, Backbone=Anole2026.06 | 66.7 | 33.3 | |
| MVoTSize=5, Backbone=Anole2026.06 | 53.3 | 36.7 | |
| T-CoTSize=5, Backbone=Anole2026.06 | 43.3 | 26.7 | |
| MVoTSize=6, Backbone=Anole2026.06 | 43.3 | 20 | |
| T-CoTSize=6, Backbone=Anole2026.06 | 33.3 | 13.3 |