Binary Classification on CounterAnimal Pair 2: Ptarmigan vs. Prairie-Chicken
84.1Average ScoreDIAL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DIALBackbone=SigLIP2, Evaluation Protocol=Zero-shot2026.05 | 84.1 | 53.9 | |
| BAP (ZS)Backbone=SigLIP2, Evaluation Protocol=Zero-shot2026.05 | 83.8 | 59.2 | |
| Native Backbone (ZS)Backbone=CLIP, Evaluation Protocol=Zero-shot2026.05 | 81.6 | 46.4 | |
| Native Backbone (ZS)Backbone=SigLIP2, Evaluation Protocol=Zero-shot2026.05 | 81.4 | 44.7 | |
| BAP (LP)Backbone=CLIP, Evaluation Protocol=Trained2026.05 | 81.4 | 65 | |
| Native Backbone (LP)Backbone=CLIP, Evaluation Protocol=Trained2026.05 | 80.5 | 50.4 | |
| BAP (LP)Backbone=SigLIP2, Evaluation Protocol=Trained2026.05 | 80.4 | 69.3 | |
| BAP (ZS)Backbone=CLIP, Evaluation Protocol=Zero-shot2026.05 | 79.1 | 55.8 | |
| DIALBackbone=CLIP, Evaluation Protocol=Zero-shot2026.05 | 76.7 | 51.3 | |
| Data-Matched ControlBackbone=SigLIP2, Evaluation Protocol=Trained2026.05 | 75.7 | 61.3 | |
| Native Backbone (LP)Backbone=SigLIP2, Evaluation Protocol=Trained2026.05 | 70.9 | 46.3 | |
| Data-Matched ControlBackbone=CLIP, Evaluation Protocol=Trained2026.05 | 67 | 43.9 | |
| ERM (LP-FT)Backbone=CLIP, Evaluation Protocol=Trained2026.05 | 65.3 | 19.2 | |
| ERM (LP-FT)Backbone=SigLIP2, Evaluation Protocol=Trained2026.05 | 61.8 | 14.7 | |
| PRuSCBackbone=SigLIP2, Evaluation Protocol=Trained2026.05 | 52.8 | 21.6 | |
| PRuSCBackbone=CLIP, Evaluation Protocol=Trained2026.05 | 47.1 | 29.7 |