Scientific Text Revision
62Pairwise Accuracy*-PLUIE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| *-PLUIEJudge Model=Llama, Reference Provided (w.g.)=true2026.02 | 62 | 0.32 | 0.34 | |
| *-PLUIEJudge Model=Phi, Reference Provided (w.g.)=false2026.02 | 61 | 0.31 | 0.32 | |
| *-PLUIEJudge Model=Phi, Reference Provided (w.g.)=true2026.02 | 61 | 0.32 | 0.33 | |
| *-PLUIEJudge Model=Llama, Reference Provided (w.g.)=false2026.02 | 61 | 0.31 | 0.32 | |
| LLM-choiceJudge Model=Llama, Reference Provided (w.g.)=true2026.02 | 60 | 0.3 | 0.31 | |
| LLM-choiceJudge Model=Llama, Reference Provided (w.g.)=false2026.02 | 59 | 0.28 | 0.3 | |
| LLM-choiceJudge Model=Phi, Reference Provided (w.g.)=true2026.02 | 55 | 0.27 | 0.27 | |
| LLM-choiceJudge Model=Phi, Reference Provided (w.g.)=false2026.02 | 53 | 0.25 | 0.24 | |
| Para-PLUIEJudge Model=Phi, Reference Provided (w.g.)=false2026.02 | 52 | 0.21 | 0.15 | |
| Para-PLUIEJudge Model=Llama, Reference Provided (w.g.)=false2026.02 | 52 | 0.2 | 0.17 | |
| LLM-LikertJudge Model=Phi, Reference Provided (w.g.)=true2026.02 | 52 | 0.29 | 0.26 | |
| LLM-LikertJudge Model=Llama, Reference Provided (w.g.)=true2026.02 | 50 | 0.27 | 0.23 | |
| LLM-LikertJudge Model=Phi, Reference Provided (w.g.)=false2026.02 | 45 | 0.3 | 0.21 | |
| BERTScoreReference Provided (w.g.)=false2026.02 | 45 | 0.16 | 0.03 | |
| LLM-LikertJudge Model=Llama, Reference Provided (w.g.)=false2026.02 | 44 | 0.29 | 0.19 | |
| LevenshteinReference Provided (w.g.)=false2026.02 | 44 | 0.13 | 0.01 | |
| METEORReference Provided (w.g.)=false2026.02 | 42 | 0.19 | 0 | |
| BLEUReference Provided (w.g.)=false2026.02 | 41 | 0.17 | -0.03 | |
| Modern BERTScoreReference Provided (w.g.)=false2026.02 | 36 | 0.13 | -0.07 | |
| Random weightedReference Provided (w.g.)=false2026.02 | 36 | 0.03 | 0 | |
| Random uniformReference Provided (w.g.)=false2026.02 | 33 | 0.03 | 0 |