LGT Detection on Fast-DetectGPT XSum (test)
100AUROCGPTZero
Evaluation Results
| Method | Links | |
|---|---|---|
| GPTZeroSource LLM=GPT4 Turbo2024.05 | 100 | |
| GPTZeroSource LLM=Llama3 70B2024.05 | 100 | |
| ReMoDetectModel=GPT4 Turbo2024.05 | 100 | |
| GPTZeroSource LLM=Claude3 Opus2024.05 | 99.9 | |
| ReMoDetectModel=GPT3.5 Turbo2024.05 | 99.9 | |
| ReMoDetectModel=Claude 3 Opus2024.05 | 99.9 | |
| ReMoDetectSource LLM=GPT 3.5 Turbo2024.05 | 99.8 | |
| ReMoDetectSource LLM=GPT4 Turbo2024.05 | 99.8 | |
| ReMoDetectModel=Llama3 70B2024.05 | 99.8 | |
| Fast-DetectGPTModel=Llama3 70B2024.05 | 99.7 | |
| ReMoDetectSource Model=Qwen2 1.5B-it2024.05 | 99.6 | |
| GPTZeroSource LLM=GPT 3.5 Turbo2024.05 | 99.5 | |
| ReMoDetectSource LLM=Llama3 70B2024.05 | 99.5 | |
| ReMoDetectSource LLM=Claude3 Opus2024.05 | 99.5 | |
| ReMoDetectSource Model=Gemma2 9B-it2024.05 | 99.5 | |
| Fast-DetectGPTSource Model=OLMo 7B-sft2024.05 | 99.1 | |
| Fast-DetectGPTModel=GPT3.5 Turbo2024.05 | 99.1 | |
| ReMoDetectSource LLM=GPT42024.05 | 98.8 | |
| ReMoDetectModel=GPT42024.05 | 98.7 | |
| Fast-DetectGPTSource Model=Qwen2 1.5B-it2024.05 | 98.3 | |
| GPTZeroSource LLM=GPT42024.05 | 98.2 | |
| ReMoDetectSource Model=OLMo 7B-sft2024.05 | 98.1 | |
| Log-likelihoodModel=Llama3 70B2024.05 | 97.9 | |
| ChatGPT-DetectorSource Model=Gemma2 9B-it2024.05 | 97.7 | |
| Fast-DetectGPTModel=GPT4 Turbo2024.05 | 97.4 | |
| OpenAI-DetectorSource Model=OLMo 7B-sft2024.05 | 97.3 | |
| ChatGPT-DetectorSource Model=Gemma2 2B-it2024.05 | 97.2 | |
| OpenAI-DetectorSource Model=Qwen2 1.5B-it2024.05 | 97.2 | |
| Log-likelihoodSource Model=OLMo 7B-sft2024.05 | 96.6 | |
| Log-likelihoodSource Model=Qwen2 1.5B-it2024.05 | 96.5 | |
| OpenAI-DetectorModel=Llama3 70B2024.05 | 96.2 | |
| Fast-DetectGPTModel=Claude 3 Opus2024.05 | 96.2 | |
| Log-likelihoodModel=Claude 3 Opus2024.05 | 95.9 | |
| Log-likelihoodModel=GPT3.5 Turbo2024.05 | 95.8 | |
| LRRModel=Llama3 70B2024.05 | 95.5 | |
| OpenAI-DetectorSource Model=Llama3 8B-it2024.05 | 95.4 | |
| ReMoDetectSource Model=Gemma2 2B-it2024.05 | 94.9 | |
| Fast-DetectGPTSource Model=Gemma2 9B-it2024.05 | 94 | |
| NPRModel=Llama3 70B2024.05 | 93.8 | |
| DetectGPTModel=Llama3 70B2024.05 | 93.2 | |
| LRRModel=GPT3.5 Turbo2024.05 | 91.6 | |
| OpenAI-DetectorModel=GPT3.5 Turbo2024.05 | 91.5 | |
| Log-likelihoodModel=GPT4 Turbo2024.05 | 90.9 | |
| Fast-DetectGPTModel=GPT42024.05 | 90.7 | |
| Fast-DetectGPTSource Model=Gemma2 2B-it2024.05 | 89.8 | |
| LRRModel=Claude 3 Opus2024.05 | 89.7 | |
| DetectGPTModel=GPT3.5 Turbo2024.05 | 89.2 | |
| OpenAI-DetectorModel=GPT4 Turbo2024.05 | 88.2 | |
| LRRModel=GPT4 Turbo2024.05 | 87.9 | |
| Fast-DetectGPTSource Model=Llama3 8B-it2024.05 | 86.8 | |
| NPRModel=GPT3.5 Turbo2024.05 | 86.6 | |
| OpenAI-DetectorModel=Claude 3 Opus2024.05 | 86.2 | |
| GPTZeroSource LLM=Gemini pro2024.05 | 85.8 | |
| ReMoDetectSource Model=Llama3 8B-it2024.05 | 85.4 | |
| DetectGPTModel=Claude 3 Opus2024.05 | 85.3 | |
| Log-likelihoodSource Model=Gemma2 9B-it2024.05 | 85.1 | |
| NPRModel=Claude 3 Opus2024.05 | 84.7 | |
| DetectGPTModel=GPT4 Turbo2024.05 | 83.2 | |
| Log-likelihoodSource Model=Llama3 8B-it2024.05 | 82.3 | |
| Log-likelihoodSource Model=Gemma2 2B-it2024.05 | 82.1 | |
| NPRModel=GPT4 Turbo2024.05 | 81.8 | |
| Log-likelihoodModel=GPT42024.05 | 79.8 | |
| NPRModel=Gemini pro2024.05 | 79.6 | |
| Fast-DetectGPTModel=Gemini pro2024.05 | 79.5 | |
| Log-likelihoodModel=Gemini pro2024.05 | 78.6 | |
| RankModel=GPT3.5 Turbo2024.05 | 74.9 | |
| RankModel=Llama3 70B2024.05 | 74.9 | |
| ReMoDetectSource LLM=Gemini pro2024.05 | 74.5 | |
| LRRModel=GPT42024.05 | 74.5 | |
| ReMoDetectModel=Gemini pro2024.05 | 74.5 | |
| OpenAI-DetectorSource Model=Gemma2 9B-it2024.05 | 74 | |
| RankModel=GPT4 Turbo2024.05 | 73.4 | |
| LRRModel=Gemini pro2024.05 | 73 | |
| DetectGPTModel=Gemini pro2024.05 | 72.8 | |
| OpenAI-DetectorModel=Gemini pro2024.05 | 72.2 | |
| RankModel=Claude 3 Opus2024.05 | 71.1 | |
| RankSource Model=Gemma2 9B-it2024.05 | 69.4 | |
| RankSource Model=Llama3 8B-it2024.05 | 68.9 | |
| OpenAI-DetectorModel=GPT42024.05 | 67.8 | |
| OpenAI-DetectorSource Model=Gemma2 2B-it2024.05 | 67.6 | |
| DetectGPTModel=GPT42024.05 | 67.1 | |
| RankSource Model=Qwen2 1.5B-it2024.05 | 66.7 | |
| RankModel=GPT42024.05 | 66.4 | |
| RankSource Model=OLMo 7B-sft2024.05 | 66 | |
| NPRModel=GPT42024.05 | 64.8 | |
| ChatGPT-DetectorModel=Gemini pro2024.05 | 54.7 | |
| ChatGPT-DetectorModel=GPT42024.05 | 50.3 | |
| RankModel=Gemini pro2024.05 | 44.5 | |
| RankSource Model=Gemma2 2B-it2024.05 | 18.2 | |
| ChatGPT-DetectorSource Model=Llama3 8B-it2024.05 | 13.1 | |
| ChatGPT-DetectorModel=GPT3.5 Turbo2024.05 | 9.7 | |
| ChatGPT-DetectorModel=Llama3 70B2024.05 | 7.1 | |
| ChatGPT-DetectorSource Model=OLMo 7B-sft2024.05 | 5.9 | |
| ChatGPT-DetectorModel=Claude 3 Opus2024.05 | 5.3 | |
| ChatGPT-DetectorModel=GPT4 Turbo2024.05 | 4.4 | |
| ChatGPT-DetectorSource Model=Qwen2 1.5B-it2024.05 | 1.3 |