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

Safety Alignment Evaluation on Strata

99Safety Rate

MESA

Updated 1d ago

Evaluation Results

Method	Links
MESA 2026.05		99
GRPO 2026.05		96
MESA 2026.05		95
Stair-SFT 2026.05		94
Stair-DPO 2026.05		93
Stair-SFT 2026.05		92.5
SFT 2026.05		92
SafeX 2026.05		91
SFT 2026.05		88.5
Base(instruct) 2026.05		88
Stair-DPO 2026.05		87.5
SafeX 2026.05		81
Base(chat) 2026.05		70.5
GRPO 2026.05		64