List your company

Bias Evaluation on GenderIdentity

0.12Mean ImprovementiPASwo
0.0160.0430.070.097Sep 25, 2025

Evaluation results

MethodLinks
iPASwoModel=DeepSeek-R1-Distill-8B, Steering Target=residual2025.090.120.10
iPASwoModel=Nous-Hermes-2-Mistral-7B-DPO, Steering Target=residual2025.090.10.090
iPASwoModel=Llama-3.1-8B-Instruct, Steering Target=residual2025.090.10.090
iPASaModel=Nous-Hermes-2-Mistral-7B-DPO, Steering Target=residual2025.090.090.080
iPASaModel=Llama-3.1-8B-Instruct, Steering Target=residual2025.090.090.080
iPASaModel=DeepSeek-R1-Distill-8B, Steering Target=residual2025.090.060.050
PASfModel=Nous-Hermes-2-Mistral-7B-DPO, Steering Target=residual2025.090.050.040
PASfModel=Llama-3.1-8B-Instruct, Steering Target=residual2025.090.040.030
PASfModel=DeepSeek-R1-Distill-8B, Steering Target=residual2025.090.020.020