Stereotype Bias Evaluation on StereoSet Gender
99.59LMS ScoreLLaMa3-8b
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LLaMa3-8bMethod Type=Original2026.05 | 99.59 | 57.7 | 84.3 | |
| originalBase Model=GPT-22026.05 | 98.76 | 54.81 | 89.26 | |
| DebiasRAGBase Model=GPT-22026.05 | 92.98 | 51.44 | 90.21 | |
| LLaMa3-8bMethod Type=Regular RAG2026.05 | 92.56 | 58.93 | 76.03 | |
| DebiasRAGBase Model=LLaMa3-8b2026.05 | 92.15 | 53.85 | 85.05 | |
| originalLanguage Model=BERT2026.05 | 86.34 | 59.66 | 69.66 | |
| DebiasRAGLanguage Model=BERT2026.05 | 86.01 | 55.62 | 76.34 | |
| Sent-DebiasBackbone=BERT2023.07 | 85.6 | 59.05 | 70.11 | |
| CDABackbone=BERT2023.07 | 85.47 | 58.69 | 70.63 | |
| DropoutBackbone=BERT2023.07 | 85.42 | 60.11 | 68.16 | |
| ELEATRABackbone=ELEATRA2023.07 | 85.12 | 58.15 | 71.24 | |
| MABELBackbone=BERT2023.07 | 84.8 | 56.92 | 73.07 | |
| DistilBERTBackbone=DistilBERT2023.07 | 84.75 | 60.52 | 66.93 | |
| ADEPTLanguage Model=BERT2026.05 | 84.65 | 56.02 | 74.46 | |
| CCPABackbone=ELEATRA, prompt_template=(1, 1, 1)2023.07 | 84.63 | 52.97 | 79.61 | |
| CCPABackbone=BERT, prompt_template=(1, 1, 1)2023.07 | 84.44 | 56.61 | 73.28 | |
| DPCELanguage Model=BERT2026.05 | 84.42 | 59.66 | 68.12 | |
| BERTBackbone=BERT2023.07 | 84.17 | 60.28 | 66.86 | |
| INLPBackbone=BERT2023.07 | 82.69 | 58.09 | 69.31 | |
| CCPABackbone=DistilBERT, prompt_template=(1, 1, 1)2023.07 | 81.91 | 56.47 | 71.3 | |
| Auto-DebiasBackbone=BERT2023.07 | 81.76 | 57.33 | 69.78 | |
| Davinci2022.05 | 75.6 | 66.5 | 50.6 | |
| OPT-175B2022.05 | 74 | 63.6 | 53.8 | |
| FairFilBackbone=BERT2023.07 | 48.78 | 46.44 | 45.31 |