Multi-Agent System Routing on MMLU
68.2AccuracyANTAP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ANTAPAttack=Sleeper2026.06 | 68.2 | 0.9 | |
| ANTAPAttack=VPI2026.06 | 68.2 | 0.9 | |
| ANTAPAttack=MTBA2026.06 | 65.9 | 1.9 | |
| ANTAPAttack=CTBA2026.06 | 65.6 | 0.3 | |
| ANTAPAttack=Description2026.06 | 65.1 | 0.2 | |
| EmbedLLMAttack=Sleeper2026.06 | 52.3 | 17 | |
| EmbedLLMAttack=MTBA2026.06 | 51.2 | 15.2 | |
| EmbedLLMAttack=VPI2026.06 | 51.1 | 12.2 | |
| EmbedLLMAttack=Description2026.06 | 46.3 | 5.8 | |
| EmbedLLMAttack=CTBA2026.06 | 45.8 | 18.5 | |
| ANTAPAttack=BadNet2026.06 | 44.9 | 2.4 | |
| AutoGenAttack=VPI2026.06 | 41.1 | 74.7 | |
| AutoGenAttack=Sleeper2026.06 | 39.3 | 73.5 | |
| AutoGenAttack=MTBA2026.06 | 37.9 | 74.5 | |
| AutoGenAttack=CTBA2026.06 | 34.3 | 78.8 | |
| EmbedLLMAttack=BadNet2026.06 | 31.8 | 21.3 | |
| AutoGenAttack=BadNet2026.06 | 29.4 | 73.2 | |
| AutoGenAttack=Description2026.06 | 12.3 | 75.3 |