Text Classification on AGNews (Accuracy and p-value)
93.6AccuracyEspeW
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EspeWAttack=Dimension-shift Attack2026.04 | 93.6 | 10 | |
| WARDENAttack=No Attack2026.04 | 93.48 | 10 | |
| GeoMarkAttack=Dimension-shift Attack2026.04 | 93.47 | 10 | |
| GeoMarkAttack=No Attack2026.04 | 93.45 | 10 | |
| OriginalAttack=None2026.04 | 93.43 | 0.99 | |
| WARDENAttack=Dimension-shift Attack2026.04 | 93.43 | 10 | |
| EspeWAttack=No Attack2026.04 | 93.42 | 10 | |
| WARDENAttack=Dimension-reduction Attack2026.04 | 93.36 | 10 | |
| WETAttack=Paraphrasing Attack (NLLB)2026.04 | 93.32 | 10 | |
| WETAttack=Dimension-shift Attack2026.04 | 93.27 | 0.92 | |
| GeoMarkAttack=Dimension-reduction Attack2026.04 | 93.24 | 10 | |
| WETAttack=No Attack2026.04 | 93.21 | 10 | |
| EspeWAttack=Dimension-reduction Attack2026.04 | 93.2 | 10 | |
| WARDENAttack=CSE Attack2026.04 | 92.99 | 10 | |
| GeoMarkAttack=CSE Attack2026.04 | 92.97 | 10 | |
| EspeWAttack=CSE Attack2026.04 | 92.93 | 10 | |
| EspeWAttack=Paraphrasing Attack (GPT-4o-mini)2026.04 | 92.88 | 10 | |
| GeoMarkAttack=Paraphrasing Attack (GPT-4o-mini)2026.04 | 92.81 | 10 | |
| WARDENAttack=Paraphrasing Attack (GPT-4o-mini)2026.04 | 92.77 | 10 | |
| WETAttack=Paraphrasing Attack (GPT-4o-mini)2026.04 | 92.66 | 10 | |
| WARDENAttack=Paraphrasing Attack (NLLB)2026.04 | 92.54 | 10 | |
| GeoMarkAttack=Paraphrasing Attack (NLLB)2026.04 | 92.5 | 10 | |
| EspeWAttack=Paraphrasing Attack (NLLB)2026.04 | 92.49 | 10 | |
| WETAttack=CSE Attack2026.04 | 90.53 | 10 | |
| ICLModel=LLaMA-2-13B2026.05 | 88.04 | — | |
| ICLModel=LLaMA-3.1-8B2026.05 | 87.16 | — | |
| LTVModel=LLaMA-2-13B2026.05 | 86.68 | — | |
| State VectorModel=LLaMA-2-13B2026.05 | 84.64 | — | |
| LTVModel=LLaMA-3.1-8B2026.05 | 82.84 | — | |
| Task VectorModel=LLaMA-3.1-8B2026.05 | 81.36 | — | |
| State VectorModel=LLaMA-3.1-8B2026.05 | 80.28 | — | |
| Task VectorModel=LLaMA-2-13B2026.05 | 79.84 | — | |
| I2CLModel=LLaMA-2-13B2026.05 | 78.88 | — | |
| I2CLModel=LLaMA-3.1-8B2026.05 | 76.76 | — | |
| Function VectorModel=LLaMA-3.1-8B2026.05 | 76.16 | — | |
| Zero-shotModel=LLaMA-3.1-8B2026.05 | 75 | — | |
| Function VectorModel=LLaMA-2-13B2026.05 | 74 | — | |
| Zero-shotModel=LLaMA-2-13B2026.05 | 73.8 | — |