List your company
ResearchFormerly WizWand

LLM-generated text detection on Xsum, WritingPrompts, and SQuAD (Gemini-1.5-Flash test set)

75.14AUROCSurpMark
43.284851.554959.82568.0951Oct 8, 2025
Updated 5mo ago

Evaluation results

MethodLinks
SurpMarkBlack-box setting=true, k (number of discretization states)=7, B (reference-per-test budget)=22025.1075.14
SurpMarkBlack-box setting=true, k (number of discretization states)=6, B (reference-per-test budget)=22025.1074.57
BinocularsBlack-box setting=true2025.1074.51
Fast-DetectGPTBlack-box setting=true2025.1072.49
Lastde++Black-box setting=true, B (reference-per-test budget)=1002025.1071.72
R-DetectBlack-box setting=true2025.1069.25
DetectGPTBlack-box setting=true, B (reference-per-test budget)=1002025.1069.19
DetectNPRBlack-box setting=true, B (reference-per-test budget)=1002025.1064.96
DNA-GPTBlack-box setting=true, B (reference-per-test budget)=102025.1062.06
FourierGPTBlack-box setting=true2025.1061.25
EntropyBlack-box setting=true2025.1058.36
LikelihoodBlack-box setting=true2025.1056.49
LogRankBlack-box setting=true2025.1053.87
LastdeBlack-box setting=true2025.1048.13
DetectLRRBlack-box setting=true2025.1044.51