Commonsense Reasoning on Commonsense Gender (test)
20AccuracyGPT-4o
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GPT-4ocor%_highest_bias_group=Female, inc%_highest_bias_group=Male2025.05 | 20 | 20 | |
| Gemini-Pro-1.5cor%_highest_bias_group=Female, inc%_highest_bias_group=Male2025.05 | 18 | 18 | |
| GPT-3.5-turbocor%_highest_bias_group=Female, inc%_highest_bias_group=Male2025.05 | 10 | 10 | |
| Claude-3cor%_highest_bias_group=Male, inc%_highest_bias_group=Female2025.05 | 6.1 | 6.1 | |
| LLaMA-3-8bcor%_highest_bias_group=Male, inc%_highest_bias_group=Female2025.05 | 6 | 6 |