Question Answering on SIQA (test)
94.06AccuracyML Routing
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ML RoutingAdversarial attack=TextAttack2026.06 | 94.06 | 94.15 | 96.81 | 3.86 | |
| LWV FusionAdversarial attack=TextAttack2026.06 | 93.64 | 93.76 | 97.92 | 2.75 | |
| Geometric-Based ModelAdversarial attack=TextAttack2026.06 | 75.51 | 72.81 | — | 22.87 | |
| FreeLBAdversarial attack=TextAttack2026.06 | 65.03 | 65.03 | — | 34.97 | |
| Uncertainty-Based ModelAdversarial attack=TextAttack2026.06 | 62.55 | 62.55 | 62.59 | 37.45 | |
| Entropy-Based ModelAdversarial attack=TextAttack2026.06 | 61.5 | 56.77 | 76.81 | 23.19 | |
| Base Model (no defence)Adversarial attack=TextAttack2026.06 | 40.6 | 50.41 | — | 63.45 |