AI-generated text detection on MixSet 1.0 (test)
99Average ScoreGigaCheck
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GigaCheckBackbone=Mistral-7B2024.10 | 99 | 99 | 99 | 99 | 99 | 99 | 99 | 99 | 99 | 99 | 99 | 99 | 99 | |
| Radar2024.10 | 87.6 | 86.7 | 87.7 | 87.7 | 87.7 | 87.7 | 87.7 | 87.7 | 87.7 | 87.7 | 87.7 | 87.7 | 87.7 | |
| GPT-sentinel2024.10 | 71.3 | 71.4 | 71.4 | 71.4 | 71.4 | 71.4 | 71.4 | 71.4 | 71.4 | 69.6 | 71.4 | 71.4 | 71.4 | |
| DistilBERT2024.10 | 66.4 | 66.7 | 66.7 | 66.7 | 66.7 | 66.7 | 66.7 | 66.7 | 66.7 | 63.9 | 66.7 | 66.7 | 66.7 | |
| Entropy2024.10 | 64.8 | 69 | 67.1 | 68.1 | 61.3 | 68.1 | 67.1 | 68.1 | 67.1 | 62.3 | 43 | 68.1 | 68.1 | |
| DetectGPT2024.10 | 63.5 | 71.5 | 65.1 | 65.6 | 56 | 63.2 | 58.7 | 65.7 | 63.2 | 69.2 | 58.7 | 64.1 | 60.9 | |
| log likelihood2024.10 | 62.4 | 69.5 | 69.5 | 63.7 | 49.2 | 65.7 | 62.7 | 65.7 | 65.7 | 63.7 | 38.6 | 67.6 | 66.7 | |
| log-rank2024.10 | 61.5 | 69.5 | 68.6 | 63.7 | 47.9 | 61.7 | 60.6 | 64.7 | 59.5 | 61.7 | 45.4 | 67.6 | 66.7 | |
| GLTR2024.10 | 58.8 | 68.6 | 64.7 | 60.6 | 44.1 | 57.4 | 58.5 | 63.7 | 54 | 61.7 | 40 | 65.7 | 66.7 |