AI-generated text detection on AIGenPoetry (test)
89.44Macro-F1 (GPT-4.1)RoBERTa
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| RoBERTa2026.05 | 89.44 | 73.31 | 78.44 | 96.74 | 84.48 | |
| Binoculars2026.05 | 34.32 | 61.73 | 69.08 | 33.83 | 49.74 | |
| LRR2026.05 | 33.83 | 39.31 | 48.18 | 33.33 | 38.66 | |
| Log-Likelihood2026.05 | 33.61 | 45.03 | 48.93 | 33.33 | 40.23 | |
| Log-Rank2026.05 | 33.61 | 43.56 | 47.17 | 33.33 | 39.42 | |
| Fast-DetectGPT2026.05 | 33.33 | 51.96 | 62.73 | 33.28 | 45.33 |