Prompt Injection on OpenPromptInjection
73.6ASVhUCC
Evaluation results
| Method | Links | ||
|---|---|---|---|
| UCCModel Type=Poison, Backbone=GEM3-12B2026.05 | 73.6 | 73.7 | |
| UCCModel Type=Poison, Backbone=QWE3-30B2026.05 | 73.3 | 73.5 | |
| CoAModel Type=Base, Backbone=GEM3-12B2026.05 | 73.2 | 74 | |
| UCCModel Type=Poison, Backbone=QWE3-4B2026.05 | 72 | 72.1 | |
| UCCModel Type=Poison, Backbone=PHI4-15B2026.05 | 71.6 | 71.7 | |
| UCCModel Type=Poison, Backbone=LLA3-8B2026.05 | 71.2 | 71.3 | |
| BadNetModel Type=Poison, Backbone=QWE3-30B2026.05 | 59.2 | 59.3 | |
| CoAModel Type=Clean, Backbone=GEM3-12B2026.05 | 58.7 | 59.8 | |
| MTBAModel Type=Poison, Backbone=QWE3-30B2026.05 | 58.6 | 58.6 | |
| CTBAModel Type=Poison, Backbone=QWE3-30B2026.05 | 58.3 | 58.3 | |
| VPIModel Type=Poison, Backbone=QWE3-30B2026.05 | 55.8 | 55.9 | |
| CoAModel Type=Clean, Backbone=QWE3-30B2026.05 | 55.3 | 56.3 | |
| CoAModel Type=Clean, Backbone=QWE3-4B2026.05 | 54.4 | 58.7 | |
| CoAModel Type=Clean, Backbone=PHI4-15B2026.05 | 51.2 | 54.5 | |
| VPIModel Type=Poison, Backbone=GEM3-12B2026.05 | 51 | 51.1 | |
| CTBAModel Type=Poison, Backbone=GEM3-12B2026.05 | 50.9 | 51 | |
| CoAModel Type=Base, Backbone=QWE3-4B2026.05 | 50.8 | 59.2 | |
| CoAModel Type=Base, Backbone=QWE3-30B2026.05 | 48.4 | 59.3 | |
| BadNetModel Type=Poison, Backbone=GEM3-12B2026.05 | 47.8 | 47.9 | |
| MTBAModel Type=Poison, Backbone=GEM3-12B2026.05 | 47.8 | 47.9 | |
| BadNetModel Type=Poison, Backbone=PHI4-15B2026.05 | 46 | 46.1 | |
| VPIModel Type=Poison, Backbone=PHI4-15B2026.05 | 45 | 45 | |
| CTBAModel Type=Poison, Backbone=PHI4-15B2026.05 | 44.9 | 44.9 | |
| MTBAModel Type=Poison, Backbone=PHI4-15B2026.05 | 44.4 | 44.4 | |
| BadNetModel Type=Poison, Backbone=LLA3-8B2026.05 | 43.8 | 43.9 | |
| MTBAModel Type=Poison, Backbone=LLA3-8B2026.05 | 42.5 | 42.7 | |
| CTBAModel Type=Poison, Backbone=LLA3-8B2026.05 | 41.1 | 41.1 | |
| VPIModel Type=Poison, Backbone=LLA3-8B2026.05 | 41 | 41.1 | |
| CoAModel Type=Clean, Backbone=LLA3-8B2026.05 | 39.2 | 45.1 | |
| SleeperModel Type=Poison, Backbone=QWE3-30B2026.05 | 38.5 | 38.6 | |
| BadNetModel Type=Poison, Backbone=QWE3-4B2026.05 | 35.9 | 36.5 | |
| CTBAModel Type=Poison, Backbone=QWE3-4B2026.05 | 33.5 | 34.1 | |
| VPIModel Type=Poison, Backbone=QWE3-4B2026.05 | 30.6 | 32.2 | |
| CoAModel Type=Base, Backbone=LLA3-8B2026.05 | 29.9 | 53.2 | |
| MTBAModel Type=Poison, Backbone=QWE3-4B2026.05 | 29.2 | 30.8 | |
| SleeperModel Type=Poison, Backbone=GEM3-12B2026.05 | 25.7 | 25.7 | |
| SleeperModel Type=Poison, Backbone=LLA3-8B2026.05 | 25.2 | 25.2 | |
| SleeperModel Type=Poison, Backbone=QWE3-4B2026.05 | 20.6 | 20.9 | |
| CoAModel Type=Base, Backbone=PHI4-15B2026.05 | 18 | 37.9 | |
| SleeperModel Type=Poison, Backbone=PHI4-15B2026.05 | 15.8 | 15.8 |