Instruction Following on AlpacaEval Yoruba
68.9Win Rate (%)CLO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CLOModel=Qwen-2.5-3B, Eval Lang=Target2025.05 | 68.9 | 23 | |
| CLOModel=Llama-2-13B, Eval Lang=Target2025.05 | 67.3 | 23.8 | |
| CLOModel=Llama-2-7B, Eval Lang=Target2025.05 | 67.1 | 23.6 | |
| CLOModel=Llama-3-8B, Eval Lang=Target2025.05 | 64 | 11.6 | |
| CLOModel=Llama-2-7B, Eval Lang=English2025.05 | 61.2 | 10.9 | |
| CLOModel=Llama-3-8B, Eval Lang=English2025.05 | 58.2 | 7 | |
| CLOModel=Qwen-2.5-3B, Eval Lang=English2025.05 | 57 | 2.6 | |
| SFT+DPOModel=Qwen-2.5-3B, Eval Lang=English2025.05 | 54.4 | — | |
| CLOModel=Llama-2-13B, Eval Lang=English2025.05 | 54.2 | 3.1 | |
| CLOModel=Mistral-7B-v0.1, Eval Lang=English2025.05 | 52.8 | 3.4 | |
| SFT+DPOModel=Llama-3-8B, Eval Lang=Target2025.05 | 52.4 | — | |
| SFT+DPOModel=Llama-3-8B, Eval Lang=English2025.05 | 51.2 | — | |
| SFT+DPOModel=Llama-2-13B, Eval Lang=English2025.05 | 51.1 | — | |
| CLOModel=Mistral-7B-v0.1, Eval Lang=Target2025.05 | 51.1 | 0.2 | |
| SFT+DPOModel=Mistral-7B-v0.1, Eval Lang=Target2025.05 | 50.9 | — | |
| SFT+DPOModel=Llama-2-7B, Eval Lang=English2025.05 | 50.3 | — | |
| SFT+DPOModel=Mistral-7B-v0.1, Eval Lang=English2025.05 | 49.4 | — | |
| SFT+DPOModel=Qwen-2.5-3B, Eval Lang=Target2025.05 | 45.9 | — | |
| SFT+DPOModel=Llama-2-7B, Eval Lang=Target2025.05 | 43.5 | — | |
| SFT+DPOModel=Llama-2-13B, Eval Lang=Target2025.05 | 43.5 | — |