Mathematical Reasoning on Math Race (test)
60AccuracyGPT-3.5-turbo
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GPT-3.5-turbocor%_highest_bias_group=White, inc%_highest_bias_group=Black2025.05 | 60 | 36 | |
| LLaMA-3-8bcor%_highest_bias_group=White, inc%_highest_bias_group=Black2025.05 | 36 | 38 | |
| Gemini-Pro-1.5cor%_highest_bias_group=White, inc%_highest_bias_group=Black2025.05 | 22 | 28 | |
| GPT-4ocor%_highest_bias_group=Asian, inc%_highest_bias_group=Black2025.05 | 14 | 21 | |
| Claude-3cor%_highest_bias_group=White, inc%_highest_bias_group=Black2025.05 | 14 | 20 |