GEC Metric Meta-evaluation on SEEDA-E Fluency Setting
0.969Pearson RIMPARA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| IMPARAMode=Reference-free2026.02 | 0.969 | 0.965 | |
| SOMEMode=Reference-free2026.02 | 0.965 | 0.965 | |
| GPT-4-EMode=Reference-free2026.02 | 0.848 | 0.987 | |
| ScribendiMode=Reference-free2026.02 | 0.826 | 0.912 | |
| GLEUReference Set=NE-Fluency2026.02 | 0.781 | 0.921 | |
| GREENReference Set=NE-Fluency2026.02 | 0.745 | 0.908 | |
| UOT-ERRANTReference Set=NE-Fluency2026.02 | 0.705 | 0.851 | |
| GREENReference Set=E-Fluency2026.02 | 0.547 | 0.802 | |
| CLEMEReference Set=NE-Fluency2026.02 | 0.473 | 0.653 | |
| UOT-ERRANTReference Set=E-Fluency2026.02 | 0.445 | 0.684 | |
| Qwen3-8B-SMode=Reference-free2026.02 | 0.418 | 0.939 | |
| GLEUReference Set=E-Fluency2026.02 | 0.278 | 0.6 | |
| PT-ERRANTReference Set=NE-Fluency2026.02 | 0.276 | 0.578 | |
| ERRANTReference Set=NE-Fluency2026.02 | 0.114 | 0.508 | |
| PT-ERRANTReference Set=E-Fluency2026.02 | 0.005 | 0.31 | |
| ERRANTReference Set=E-Fluency2026.02 | -0.005 | 0.424 | |
| CLEMEReference Set=E-Fluency2026.02 | -0.043 | 0.297 |