Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Detection of LLM-Generated Text on XSum GPT-J-6B

88.35AUROC

SurpMark ref-attack

69.047674.058879.0784.0812Oct 8, 2025
Updated 21d ago

Evaluation Results

MethodLinks
2025.10
88.35
2025.10
88.35
2025.10
88.35
2025.10
86.37
2025.10
85.38
2025.10
85.23
2025.10
84.86
2025.10
84.44
2025.10
83.85
2025.10
83.84
2025.10
78.6
2025.10
75.44
2025.10
74.09
2025.10
73.55
2025.10
69.79