Gender bias evaluation on RealWorldQuestioning Investment Recommendations 1.0 (Entire Dataset)
77.37Male More InformationLlama-3
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Llama-3Judge Model=ChatGPT-4o, Iterations=12025.05 | 77.37 | 22.62 | 0 | 0.08 | 0 | |
| DeepSeek-R1Judge Model=ChatGPT-4o, Iterations=12025.05 | 58.69 | 40.57 | 0.72 | 0.48 | 0.003 | |
| ChatGPT-3.5-turboJudge Model=ChatGPT-4o, Iterations=12025.05 | 55.47 | 39.41 | 5.1 | 0.52 | 0.01 | |
| ChatGPT-4-turboJudge Model=ChatGPT-4o, Iterations=12025.05 | 51.82 | 42.33 | 5.83 | 0.68 | 0.14 |