NLG Evaluation on HelpSteer2
0.65Spearman CorrelationQwen
Evaluation Results
| Method | Links | |
|---|---|---|
| QwenPrompting=Zero-shot2026.02 | 0.65 | |
| QwenPrompting=Few-shot2026.02 | 0.63 | |
| MixtralPrompting=Few-shot2026.02 | 0.61 | |
| GPT-4o-miniPrompting=Few-shot2026.02 | 0.58 | |
| GPT-4oPrompting=Few-shot2026.02 | 0.56 | |
| MixtralPrompting=Zero-shot2026.02 | 0.5 | |
| GPT-4oPrompting=Zero-shot2026.02 | 0.49 | |
| GPT-4o-miniPrompting=Zero-shot2026.02 | 0.47 | |
| LlamaPrompting=Few-shot2026.02 | 0.43 | |
| LlamaPrompting=Zero-shot2026.02 | 0.31 |