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

Post-training Safety and Utility Alignment on Mistral-7B

0Unsafe Rate (%)

CB

-0.9125.24411.417.556Apr 17, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
01.041002.838.161870
2026.04
09.9559.78.047.33-231.5
2026.04
1.177.13467.737.131869.9
2026.04
1.677.12467.556.961870.3
2026.04
11.26.1736.32.972.71869
2026.04
177.4723.77.887.953669.5
2026.04
22.86.3822.77.718.07-70.4