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

Safety Evaluation on XSTest (OverRefusal Score)

99.6OverRefusal Score

SafeChain

Updated 22d ago

Evaluation Results

Method	Links
SafeChain 2026.05		99.6
Base 2026.05		99.2
SInternal 2026.05		99.2
Base 2026.05		98.8
SInternal 2026.05		98.4
SafeChain 2026.05		96.4
STAR-1 2026.05		96
SafeChain 2026.05		94.4
SInternal 2026.05		94.3
Base 2026.05		93.6
STAR-1 2026.05		93.6
STAR-1 2026.05		89.6