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

cage

Benchmarks

Task NameDataset NameSOTA ResultTrend
Agent Return VolatilityCAGE-2 v1 (test)
SD2
40
Single-arm motion planningcage single-arm planning scene
Success Rate96
9
unconstrained single-arm planningcage scenario
Success Rate95
9
Red Teaming EvaluationCAGE 200 episodes 4
AHD3.74
6
Cyber Red TeamingCAGE 4 (200 episodes)
AHD9.86
4
Showing 5 of 5 rows