AI-generated text detection on XSum Generated by ChatGPT (test)
1AUROCDPIC
Evaluation Results
| Method | Links | |
|---|---|---|
| DPICAuxiliary Model=ChatGPT2023.05 | 1 | |
| DPIC(ChatGPT)Interference=Original2023.05 | 1 | |
| DPIC(ChatGPT)Interference=DIPPER2023.05 | 1 | |
| DPIC(Vicuna-7b)Interference=DIPPER2023.05 | 0.998 | |
| DPICAuxiliary Model=Vicuna-7b2023.05 | 0.9976 | |
| DPIC(Vicuna-7b)Interference=Original2023.05 | 0.9976 | |
| RADARStrategy=Supervised2023.05 | 0.9972 | |
| RADARInterference=Original2023.05 | 0.9972 | |
| DPIC(ChatGPT)Interference=Back-translate2023.05 | 0.9972 | |
| RADARInterference=DIPPER2023.05 | 0.9964 | |
| Fast-DetectGPTStrategy=Zero-shot2023.05 | 0.9907 | |
| Fast-DetectGPTInterference=Original2023.05 | 0.9907 | |
| DPIC(Vicuna-7b)Interference=Back-translate2023.05 | 0.9889 | |
| RADARInterference=Back-translate2023.05 | 0.9801 | |
| Fast-DetectGPTInterference=Back-translate2023.05 | 0.9711 | |
| LogRankStrategy=Zero-shot2023.05 | 0.9584 | |
| LogRankInterference=Original2023.05 | 0.9584 | |
| LikelihoodStrategy=Zero-shot2023.05 | 0.9577 | |
| LikelihoodInterference=Original2023.05 | 0.9577 | |
| Fast-DetectGPTInterference=DIPPER2023.05 | 0.9536 | |
| LikelihoodInterference=Back-translate2023.05 | 0.9306 | |
| LogRankInterference=Back-translate2023.05 | 0.926 | |
| LRRStrategy=Zero-shot2023.05 | 0.9164 | |
| LRRInterference=Original2023.05 | 0.9164 | |
| RoBERTa-baseBackbone=RoBERTa-base2023.05 | 0.915 | |
| ROBERTa-baseInterference=Original2023.05 | 0.915 | |
| DNA-GPTAuxiliary Model=Neo-2.72023.05 | 0.904 | |
| DNA-GPT(Neo-2.7)Interference=Original2023.05 | 0.904 | |
| DNA-GPT(Neo-2.7)Interference=Back-translate2023.05 | 0.8624 | |
| LRRInterference=Back-translate2023.05 | 0.8621 | |
| LogRankInterference=DIPPER2023.05 | 0.8596 | |
| RoBERTa-largeBackbone=RoBERTa-large2023.05 | 0.8507 | |
| ROBERTa-largeInterference=Original2023.05 | 0.8507 | |
| LRRInterference=DIPPER2023.05 | 0.8448 | |
| LikelihoodInterference=DIPPER2023.05 | 0.8438 | |
| DNA-GPTAuxiliary Model=ChatGPT2023.05 | 0.8396 | |
| DNA-GPT(ChatGPT)Interference=Original2023.05 | 0.8396 | |
| ROBERTa-baseInterference=Back-translate2023.05 | 0.8379 | |
| ROBERTa-baseInterference=DIPPER2023.05 | 0.8148 | |
| NPRInterference=Back-translate2023.05 | 0.805 | |
| DNA-GPT(ChatGPT)Interference=Back-translate2023.05 | 0.7975 | |
| DNA-GPT(ChatGPT)Interference=DIPPER2023.05 | 0.791 | |
| ROBERTa-largeInterference=DIPPER2023.05 | 0.7884 | |
| NPRStrategy=Zero-shot2023.05 | 0.7845 | |
| NPRInterference=Original2023.05 | 0.7845 | |
| DNA-GPT(Neo-2.7)Interference=DIPPER2023.05 | 0.7733 | |
| DNA-GPTAuxiliary Model=Vicuna-7b2023.05 | 0.6992 | |
| DNA-GPT(Vicuna-7b)Interference=Original2023.05 | 0.6992 | |
| ROBERTa-largeInterference=Back-translate2023.05 | 0.6853 | |
| DNA-GPT(Vicuna-7b)Interference=DIPPER2023.05 | 0.6528 | |
| DNA-GPT(Vicuna-7b)Interference=Back-translate2023.05 | 0.6175 | |
| NPRInterference=DIPPER2023.05 | 0.5648 | |
| DetectGPTInterference=Back-translate2023.05 | 0.5417 | |
| DetectGPTStrategy=Zero-shot2023.05 | 0.4594 | |
| DetectGPTInterference=Original2023.05 | 0.4594 | |
| EntropyInterference=DIPPER2023.05 | 0.4514 | |
| EntropyStrategy=Zero-shot2023.05 | 0.3305 | |
| EntropyInterference=Original2023.05 | 0.3305 | |
| DetectGPTInterference=DIPPER2023.05 | 0.3074 | |
| EntropyInterference=Back-translate2023.05 | 0.3008 |