AI-generated text detection on DetectRL Multi-LLM
97.17AUROCLAPD
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LAPDStandardization=true, Base Model=Llama2-7B, Aligned Model=Llama2-7B-Instruct2026.04 | 97.17 | — | |
| LAPD2026.04 | 97.17 | — | |
| RAIStandardization=false, Base Model=Llama2-7B, Aligned Model=Llama2-7B-Instruct2026.04 | 93.13 | — | |
| Exons-DetectTraining paradigm=Training-free2026.03 | 90.67 | 84.95 | |
| ReMoDetect2026.04 | 89.11 | — | |
| DNA-DetectLLMStandardization=true2026.04 | 89.01 | — | |
| DNA-DetectLLMTraining paradigm=Training-free2026.03 | 88.97 | 84.85 | |
| BinocularsStandardization=false2026.04 | 83.26 | — | |
| ImBD2026.04 | 83.25 | — | |
| BinocularsTraining paradigm=Training-free2026.03 | 83.21 | 82.87 | |
| IRMTraining paradigm=Training-free2026.03 | 83.02 | 76.46 | |
| Fast-DetectGPTTraining paradigm=Training-free2026.03 | 82.26 | 75.93 | |
| Fast-DetectGPTStandardization=true2026.04 | 82.13 | — | |
| BiscopeTraining paradigm=Training-based2026.03 | 79.97 | 73.2 | |
| OpenAI-DTraining paradigm=Training-based2026.03 | 78.15 | 71.9 | |
| OpenAI Detector2026.04 | 78.15 | — | |
| Lastde++Training paradigm=Training-free2026.03 | 75.36 | 69.24 | |
| Lastde++Standardization=true2026.04 | 75.18 | — | |
| R-DetectTraining paradigm=Training-based2026.03 | 67.4 | 66.56 | |
| LogRankTraining paradigm=Training-free2026.03 | 67.3 | 66.71 | |
| LogRankStandardization=false2026.04 | 67.29 | — | |
| LikelihoodTraining paradigm=Training-free2026.03 | 66.82 | 66.71 | |
| LikelihoodStandardization=false2026.04 | 66.82 | — | |
| EntropyStandardization=false2026.04 | 64.37 | — | |
| EntropyTraining paradigm=Training-free2026.03 | 64.3 | 71.92 | |
| DetectGPTTraining paradigm=Training-free2026.03 | 49.57 | 66.67 | |
| DetectGPTStandardization=true2026.04 | 43.13 | — | |
| Fast-DetectGPTTraining Source=GPT-3.52026.06 | — | 59.55 | |
| Fast-DetectGPTTraining Source=PaLM-22026.06 | — | 55.77 | |
| Fast-DetectGPTTraining Source=Claude2026.06 | — | 0.19 | |
| Fast-DetectGPTTraining Source=Llama-22026.06 | — | 56.28 | |
| RoBERTa-BaseTraining Source=GPT-3.52026.06 | — | 99.97 | |
| RoBERTa-BaseTraining Source=PaLM-22026.06 | — | 99.25 | |
| RoBERTa-BaseTraining Source=Claude2026.06 | — | 96.83 | |
| RoBERTa-BaseTraining Source=Llama-22026.06 | — | 99.45 | |
| SV-DetectTraining Source=GPT-3.52026.06 | — | 99.6 | |
| SV-DetectTraining Source=PaLM-22026.06 | — | 99.6 | |
| SV-DetectTraining Source=Claude2026.06 | — | 99.45 | |
| SV-DetectTraining Source=Llama-22026.06 | — | 99.55 |