AI-generated text detection on Cross-genre AIGT Overall (test)
86.75OASeqXGPT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SeqXGPTAdversarial attack=PWWS2024.06 | 86.75 | 32.75 | 62.25 | 4,244.66 | |
| FloodingAdversarial attack=PWWS2024.06 | 85 | 29 | 65.88 | 4,090.51 | |
| RDropAdversarial attack=PWWS2024.06 | 84 | 37.5 | 55.36 | 4,564.72 | |
| DeBERTaAdversarial attack=PWWS2024.06 | 83.75 | 26.75 | 68.06 | 4,046.65 | |
| RoBERTaAdversarial attack=PWWS2024.06 | 82.5 | 29.5 | 64.24 | 4,088.74 | |
| SCRNAdversarial attack=PWWS2024.06 | 82.5 | 62.75 | 23.94 | 5,072.48 | |
| Log-RankAdversarial attack=PWWS2024.06 | 81 | 47.75 | 41.05 | 4,394.73 | |
| Log-LikelihoodAdversarial attack=PWWS2024.06 | 79.75 | 48.25 | 39.5 | 4,388.86 | |
| RanMASKAdversarial attack=PWWS2024.06 | 77 | 38.5 | 50 | 4,050.5 | |
| BERTAdversarial attack=PWWS2024.06 | 76.25 | 37.5 | 50.82 | 4,156.03 | |
| EntropyAdversarial attack=PWWS2024.06 | 75.75 | 17 | 77.56 | 4,067.8 | |
| ChatGPT-DetectorAdversarial attack=PWWS2024.06 | 75.75 | 36.5 | 51.82 | 4,217.32 | |
| RMLMAdversarial attack=PWWS2024.06 | 75.25 | 41 | 45.51 | 4,419.3 | |
| GLTRAdversarial attack=PWWS2024.06 | 74 | 33.75 | 54.39 | 4,086.42 |