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

Adversarial Robustness against Indirect Prompt Injection on AgentDojo Average across attacks

13.18UA

PIGuard

11.042825.468939.89554.3211Mar 11, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.03
13.180.85
2026.03
15.623.63
2026.03
18.170.59
2026.03
24.553.89
2026.03
32.0613.95
2026.03
34.589.63
2026.03
34.70.26
2026.03
36.4211.82
2026.03
39.260
2026.03
41.5818.26
2026.03
41.5811.2
2026.03
42.520
2026.03
44.137.46
2026.03
45.4910.46
2026.03
48.020
2026.03
48.515.19
2026.03
51.560.54
2026.03
52.1814.06
2026.03
52.320
2026.03
52.582.21
2026.03
55.271.05
2026.03
66.613.43