List your company

Stereotype Bias Evaluation on StereoSet (test)

77.12Gender SSAuto-Debias
41.323250.616659.9169.2034Feb 4, 2026

Evaluation results

MethodLinks
Auto-DebiasBackbone=Mistral-v0.32026.0277.1292.1942.1961.591.2770.2770.0891.654.816094.9475.95
IG2Backbone=Mistral-v0.32026.0276.0394.5345.3145.9666.9461.5451.8168.1565.6851.7970.8968.35
BaseBackbone=Mistral-v0.32026.0272.3696.0953.126092.5274.0169.3993.5857.2860.8193.6773.42
Prefix PromptingBackbone=Mistral-v0.32026.0272.2792.9751.5655.191.6882.3364.2792.5966.1752.8688.6183.54
FBABackbone=Mistral-v0.32026.0268.3393.7559.3855.1189.480.2565.894.5764.6955.493.6783.54
BBABackbone=Mistral-v0.32026.0268.3393.7559.3854.1984.4177.3459.0885.6870.1254.9389.8781.01
DDPBackbone=Mistral-v0.32026.0266.3995.3164.0658.6592.5276.566.9394.0762.2263.2486.0863.29
Self-DebiasingBackbone=Mistral-v0.32026.0242.769.5359.3748.0970.8968.1956.5172.162.7255.9374.6865.82