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

LLM Alignment on Harmlessness

87.85WR

AlignX

22.90239.763556.62573.4865Sep 26, 2025Oct 18, 2025Nov 9, 2025Dec 2, 2025Dec 24, 2025Jan 15, 2026Feb 7, 2026
Updated 15d ago

Evaluation Results

MethodLinks
2026.02
87.8533.1580.6545.12
2025.09
86.216.283.851.2
2026.02
86.0533.978.943.68
2026.02
83.1834.676.0541.54
2026.02
81.523.180.1746.19
2026.02
80.223.2576.8544.6
2025.09
64.415.525.524.8
2026.02
59.863332.0319.63
2025.09
51.412.319.219.4
2025.09
49.35.333.525.8
2025.09
41.118.623.915.5
2025.09
40.616.823.115.6
2025.09
39.517.923.415
2025.09
37.620.422.113.1
2025.09
36.616.517.712.6
2025.09
36.114.817.412.9
2025.09
3515.817.112.1
2025.09
33.820.621.811.7
2025.09
33.318.316.410.5
2025.09
32.919.721.111.4
2025.09
32.419.542.718.5
2025.09
31.91842.118.7
2025.09
30.918.941.317.8
2025.09
30.722.519.69.3
2025.09
29.721.240.516.3
2026.02
25.47.2-6.06
2025.09
25.47.220.913