Paraphrasing on QQP (Semantic Metrics)
90.26Semantic FaithfulnessGPT-4o-mini + SENSE
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GPT-4o-mini + SENSEPrompting=SENSE2024.09 | 90.26 | — | — | — | 34 | 8.08 | |
| GPT-4o-miniPrompting=Vanilla2024.09 | 89.71 | — | — | — | 39 | 7.25 | |
| NeoDiffb=102025.05 | 87.42 | 91.87 | 88.79 | 45.83 | — | — | |
| GPT-3.5-turboPrompting=Vanilla2024.09 | 85.79 | — | — | — | 46.37 | 8.76 | |
| GPT-3.5-turbo + SENSEPrompting=SENSE2024.09 | 85.79 | — | — | — | 25.33 | 10.24 | |
| LLaMA3-70B + SENSEPrompting=SENSE2024.09 | 84.02 | — | — | — | 29 | 11.51 | |
| Transformerb=102025.05 | 83.93 | 94.64 | 86.02 | 54.55 | — | — | |
| LLaMA3-70BPrompting=Vanilla2024.09 | 83.71 | — | — | — | 40 | 10.85 | |
| Transformerb=52025.05 | 83.7 | 94.73 | 86.05 | 54.52 | — | — | |
| Transformerb=12025.05 | 83.64 | 92.56 | 84.96 | 57.19 | — | — | |
| CMLMb=102025.05 | 72.86 | 81.99 | 75.6 | 55.86 | — | — |