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

Scenario Generation on SafeBench CARLA v3 (scenario-level)

50.7Attack Success Rate

ChatScene (CS) + EVODRIVE

18.4626.8335.243.57Jun 2, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
50.792.493.445.7143.5
2026.06
39.472.981.251.6-
2026.06
38.261.579.151.7142.5
2026.06
3360.772.954.8255.7
2026.06
30.192.771.357.873.7
2026.06
26.485.666.460.2-
2026.06
2654.463.158.9-
2026.06
19.751.453.763.3-