Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Poisoning Defense Evaluation on Sleeper Agent 7x7 Target-Injected Tasks Poisoned

19.8Delta Attack Effectiveness

CROW

1.3926.17110.9515.729May 26, 2026
Updated 7d ago

Evaluation Results

MethodLinks
2026.05
19.8
2026.05
3.53
2026.05
2.77
2026.05
2.2
2026.05
2.1