Machine-Generated Text Detection on Average M4 benchmark (test)
100AUROCRoBERTa
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RoBERTaGenerator=ChatGPT2025.02 | 100 | 68.9 | |
| RoBERTaGenerator=GPT-42025.02 | 100 | 75.1 | |
| RoBERTaGenerator=Dolly-v22025.02 | 100 | 66.4 | |
| ExaGPTGenerator=ChatGPT2025.02 | 99.2 | 92.7 | |
| ExaGPTGenerator=GPT-42025.02 | 99 | 92.3 | |
| LR-GLTRGenerator=GPT-42025.02 | 98.1 | 89.8 | |
| LR-GLTRGenerator=ChatGPT2025.02 | 97.9 | 84.4 | |
| DNA-GPTGenerator=ChatGPT2025.02 | 91.4 | 66.4 | |
| ExaGPTGenerator=Dolly-v22025.02 | 90.4 | 70.8 | |
| LR-GLTRGenerator=Dolly-v22025.02 | 85.3 | 67.5 | |
| DNA-GPTGenerator=Dolly-v22025.02 | 72 | 66.9 | |
| DNA-GPTGenerator=GPT-42025.02 | 57.3 | 55.3 |