Text Generation on Human Judgment
0.78Spearman Correlation (ρ)QQJ
Evaluation results
| Method | Links | |
|---|---|---|
| QQJ2026.05 | 0.78 | |
| G-Eval2026.05 | 0.63 | |
| LLM Reviewer2026.05 | 0.61 | |
| MT-Bench2026.05 | 0.58 | |
| BLEU / ROUGE2026.05 | 0.31 |
| Method | Links | |
|---|---|---|
| QQJ2026.05 | 0.78 | |
| G-Eval2026.05 | 0.63 | |
| LLM Reviewer2026.05 | 0.61 | |
| MT-Bench2026.05 | 0.58 | |
| BLEU / ROUGE2026.05 | 0.31 |