List your company

BallCircle on Bullet-Safety-Gym (Normalized Metrics)

89Normalized RewardReference policy (πref)
15.1634.3353.572.67Apr 29, 2026May 2, 2026May 6, 2026May 9, 2026May 13, 2026May 16, 2026May 20, 2026

Evaluation results

MethodLinks
Reference policy (πref)2026.05893.73
BC2026.05720.7
PREFINE2026.05710.35
PPL2026.05651.82
SafeDICE2026.05590.82
FISOR2026.04340
CDT+SASAlignment=SAS, Base Model=CDT2026.043147
CPQ2026.05180.21