List your company

Safety Alignment on Harmful Dataset (test)

81Harmful ScoreNon-Aligned
32.32844.96457.670.236Feb 2, 2024

Evaluation results

MethodLinks
Non-AlignedHarmful ratio (p)=0.05, Sample number (n)=10002024.0281
Non-AlignedHarmful ratio (p)=0.2, Sample number (n)=10002024.0279.2
Non-AlignedHarmful ratio (p)=0.1, Sample number (n)=10002024.0277.6
Non-AlignedHarmful ratio (p)=0.01, Sample number (n)=10002024.0265.6
SFTHarmful ratio (p)=0.2, Sample number (n)=10002024.0260
KLHarmful ratio (p)=0.2, Sample number (n)=10002024.0256.6
VaccineHarmful ratio (p)=0.2, Sample number (n)=10002024.0256.6
SFTHarmful ratio (p)=0.1, Sample number (n)=10002024.0255.2
KLHarmful ratio (p)=0.05, Sample number (n)=10002024.0255.2
VlguardHarmful ratio (p)=0.1, Sample number (n)=10002024.0254.4
KLHarmful ratio (p)=clean, Sample number (n)=10002024.0254.4
VlguardHarmful ratio (p)=0.05, Sample number (n)=10002024.0254
KLHarmful ratio (p)=0.1, Sample number (n)=10002024.0254
VlguardHarmful ratio (p)=0.2, Sample number (n)=10002024.0253.6
KLHarmful ratio (p)=0.01, Sample number (n)=10002024.0253.6
SFTHarmful ratio (p)=0.05, Sample number (n)=10002024.0252.6
EWCHarmful ratio (p)=clean, Sample number (n)=10002024.0250.6
EWCHarmful ratio (p)=0.01, Sample number (n)=10002024.0250.6
EWCHarmful ratio (p)=0.05, Sample number (n)=10002024.0250.6
EWCHarmful ratio (p)=0.1, Sample number (n)=10002024.0250.6
EWCHarmful ratio (p)=0.2, Sample number (n)=10002024.0250.6
VlguardHarmful ratio (p)=0.01, Sample number (n)=10002024.0250
SFTHarmful ratio (p)=0.01, Sample number (n)=10002024.0249.8
VlguardHarmful ratio (p)=clean, Sample number (n)=10002024.0249.4
SFTHarmful ratio (p)=clean, Sample number (n)=10002024.0248.6
VaccineHarmful ratio (p)=0.1, Sample number (n)=10002024.0248.2
VaccineHarmful ratio (p)=0.05, Sample number (n)=10002024.0242.8
VaccineHarmful ratio (p)=clean, Sample number (n)=10002024.0242.4
VaccineHarmful ratio (p)=0.01, Sample number (n)=10002024.0242.2
Non-AlignedHarmful ratio (p)=clean, Sample number (n)=10002024.0234.2