Machine Generated Text Detection on MGT benchmark Reuters
100AUROCFast-DetectGPT
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Fast-DetectGPTSource Model=Llama3 70B2024.05 | 100 | — | — | — | — | — | |
| ReMoDetectSource Model=Llama3 70B2024.05 | 100 | — | — | — | — | — | |
| Fast-DetectGPTSource Model=Gemini pro2024.05 | 100 | — | — | — | — | — | |
| ReMoDetectSource Model=Gemini pro2024.05 | 100 | — | — | — | — | — | |
| Fast-DetectGPTSource Model=GPT3.5 Turbo2024.05 | 99.9 | — | — | — | — | — | |
| ReMoDetectSource Model=GPT3.5 Turbo2024.05 | 99.9 | — | — | — | — | — | |
| ReMoDetectSource Model=GPT4 Turbo2024.05 | 99.9 | — | — | — | — | — | |
| Loglik.Source Model=Llama3 70B2024.05 | 99.9 | — | — | — | — | — | |
| Loglik.Source Model=Gemini pro2024.05 | 99.9 | — | — | — | — | — | |
| ReMoDetectSource Model=Claude2024.05 | 99.8 | — | — | — | — | — | |
| Fast-DetectGPTSource Model=GPT4 Turbo2024.05 | 99.4 | — | — | — | — | — | |
| LRRSource Model=Gemini pro2024.05 | 99.3 | — | — | — | — | — | |
| LRRSource Model=GPT3.5 Turbo2024.05 | 98.7 | — | — | — | — | — | |
| LRRSource Model=Llama3 70B2024.05 | 98.7 | — | — | — | — | — | |
| OpenAI-DetectorSource Model=GPT3.5 Turbo2024.05 | 98.5 | — | — | — | — | — | |
| Loglik.Source Model=GPT3.5 Turbo2024.05 | 98.2 | — | — | — | — | — | |
| ChatGPT-DetectorSource Model=GPT3.5 Turbo2024.05 | 97.2 | — | — | — | — | — | |
| OpenAI-DetectorSource Model=Llama3 70B2024.05 | 96.7 | — | — | — | — | — | |
| Loglik.Source Model=GPT4 Turbo2024.05 | 95.8 | — | — | — | — | — | |
| OpenAI-DetectorSource Model=Gemini pro2024.05 | 95.3 | — | — | — | — | — | |
| LRRSource Model=GPT4 Turbo2024.05 | 94.9 | — | — | — | — | — | |
| RankSource Model=GPT3.5 Turbo2024.05 | 94.8 | — | — | — | — | — | |
| RankSource Model=GPT4 Turbo2024.05 | 93.1 | — | — | — | — | — | |
| ChatGPT-DetectorSource Model=GPT4 Turbo2024.05 | 92.4 | — | — | — | — | — | |
| ChatGPT-DetectorSource Model=Gemini pro2024.05 | 91.5 | — | — | — | — | — | |
| Loglik.Source Model=Claude2024.05 | 91.3 | — | — | — | — | — | |
| ChatGPT-DetectorSource Model=Llama3 70B2024.05 | 90.8 | — | — | — | — | — | |
| RankSource Model=Llama3 70B2024.05 | 89.7 | — | — | — | — | — | |
| Fast-DetectGPTSource Model=Claude2024.05 | 87.8 | — | — | — | — | — | |
| OpenAI-DetectorSource Model=GPT4 Turbo2024.05 | 87.5 | — | — | — | — | — | |
| RankSource Model=Gemini pro2024.05 | 83.1 | — | — | — | — | — | |
| RankSource Model=Claude2024.05 | 79.5 | — | — | — | — | — | |
| LRRSource Model=Claude2024.05 | 79.2 | — | — | — | — | — | |
| NPRSource Model=Gemini pro2024.05 | 74.9 | — | — | — | — | — | |
| DetectGPTSource Model=Gemini pro2024.05 | 73 | — | — | — | — | — | |
| NPRSource Model=Claude2024.05 | 68.7 | — | — | — | — | — | |
| DetectGPTSource Model=Claude2024.05 | 68.1 | — | — | — | — | — | |
| OpenAI-DetectorSource Model=Claude2024.05 | 65.5 | — | — | — | — | — | |
| NPRSource Model=Llama3 70B2024.05 | 59.2 | — | — | — | — | — | |
| DetectGPTSource Model=Llama3 70B2024.05 | 58.9 | — | — | — | — | — | |
| NPRSource Model=GPT4 Turbo2024.05 | 53.3 | — | — | — | — | — | |
| DetectGPTSource Model=GPT4 Turbo2024.05 | 52.6 | — | — | — | — | — | |
| NPRSource Model=GPT3.5 Turbo2024.05 | 51.1 | — | — | — | — | — | |
| DetectGPTSource Model=GPT3.5 Turbo2024.05 | 50.5 | — | — | — | — | — | |
| ChatGPT-DetectorSource Model=Claude2024.05 | 25.6 | — | — | — | — | — | |
| Entropy2024.05 | — | 77.8 | 75.5 | 78.6 | 78.3 | 77.9 | |
| GLTR2024.05 | — | 98.4 | 94.8 | 99.5 | 99.6 | 84.7 | |
| Log Rank2024.05 | — | 98.6 | 95.8 | 99.7 | 99.7 | 85.5 | |
| ReMoDetect2024.05 | — | 99.9 | 99.9 | 100 | 100 | 99.8 |