AI-generated text detection

Benchmarks

Dataset Name	SOTA Method	Metric
XSum Generated by Claude3 (test)	DPIC	AUROC100	60	4mo ago
XSum Generated by GPT-4 (test)	DPIC	AUROC0.9996	60	4mo ago
XSum Generated by ChatGPT (test)	DPIC	AUROC1	60	4mo ago
READ (test)	READER	Accuracy95.3	55	2mo ago
Long-form QA 3K generations corpus	Watermark	Detection Accuracy (1% FPR)100	42	4mo ago
RAID (test)	Binoculars	TP @ 20% Error Threshold78.4	42	4mo ago
M4	LUSR	AUROC93	41	1mo ago
AcademicResearch	Learn-to-Distance	AUC99.5	36	4mo ago
Essay	FourierGPT-Mult	AUROC (GPT4All)99.99	35	2mo ago
Cross-genre (test)	Log-Likelihood	OA97.5	32	4mo ago
RealDet	LAPD	AUROC95.32	27	3mo ago
DetectRL Multi-Domain	RAI	AUROC96.19	27	1mo ago
DetectRL Multi-LLM	LAPD	AUROC97.17	27	1mo ago
mixed-source AI -> Human GPT-2, GPT-Neo, GPT-J, LLaMa, GPT-3	SeqXGPT	Overall Accuracy96.5	26	4mo ago
TravelTourism	Learn-to-Distance	AUC99.8	24	4mo ago
TechnicalWriting	Learn-to-Distance	AUC100	24	4mo ago
Sports	Learn-to-Distance	AUC0.965	24	4mo ago
Religious	Learn-to-Distance	AUC99.6	24	4mo ago
ProductReview	Learn-to-Distance	AUC99.5	24	4mo ago
PersonalCommunication	Learn-to-Distance	AUC0.95	24	4mo ago
OnlineContent	Learn-to-Distance	AUC97.3	24	4mo ago
NewsArticle	Learn-to-Distance	AUC99.9	24	4mo ago
MedicalText	Binoculars	AUC0.985	24	4mo ago
LiteratureCreativeWriting	Learn-to-Distance	AUC100	24	4mo ago
LegalDocument	Learn-to-Distance	AUC1	24	4mo ago

Showing 25 of 191 rows

...