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

Text Moderation on HarmBench n = 400

42Flagged Count

Gemma 3-1B

-1.689.662132.34Apr 28, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
4210.23.184.6
2026.04
257.20.685.2
2026.04
226.51.794.3
2026.04
175.50.470.6
2026.04
164.60.494.7
2026.04
146.22.962.3
2026.04
133.50.181.1
2026.04
85.3244.8
2026.04
851.752.3
2026.04
01.50.39.8
2026.04
02.519.6
2026.04
00.90.110
2026.04
020.69.7