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

Adversarial safety and pedagogical evaluation on SHAPE

-2.72Delta Safe

SocraticLM

-103.1528-77.0789-51.005-24.9311Apr 24, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
-2.72-63.49
2026.04
-8.84-18.29
2026.04
-16.86-7.85
2026.04
-31.59-38.23
2026.04
-58.67-1.07
2026.04
-61.76-68.91
2026.04
-62.23-0.76
2026.04
-69.593.36
2026.04
-83-6.49
2026.04
-87.18-7.29
2026.04
-87.653.17
2026.04
-90.26-73.95
2026.04
-94.77-52.8
2026.04
-99.29-70.33