Harmfulness Evaluation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
5Harmfulness Score
56
Apr 24, 2026
5Harmfulness Score (Llama-Guard-3B)
56
Apr 24, 2026
91JBB
36
Apr 23, 2026
1.04Harmfulness Score
28
Jun 16, 2026
1.12Harmful Response Ratio
27
Jun 16, 2026
3.78Harmfulness Score
22
May 7, 2026
1.32Harmfulness Score
22
May 7, 2026
1.08Harmfulness Score
22
May 21, 2026
1.16Harmfulness Score
22
May 7, 2026
26.35ASR Score
15
May 8, 2026
78.41ASR Accuracy
15
May 8, 2026
90.8ASR
15
May 8, 2026
49ASR
12
May 8, 2026
3.25Beaver-7B Cost Score
10
Feb 26, 2026
-1.11Beaver-7B-Cost Score
10
Feb 26, 2026
4.95Average Score (Full)
6
Jun 16, 2026
1.15Primary Average
6
Jun 16, 2026
1.13Harmfulness Score
6
Jun 16, 2026
1.26Harmfulness Score
6
Jun 16, 2026
0Harmfulness Rate
6
Jun 2, 2026
0Rate
6
Jun 2, 2026
0Harmfulness Rate
6
Jun 2, 2026
0Harmful Rate
6
Jun 2, 2026
62.42HS Score
3
Jun 5, 2026
89.6Score (Effectiveness)
2
Mar 19, 2026