Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Machine-generated text detection on MAGE hswag held-out
Loading...
93.2
AUROC
confidence-weighted ensemble
86.752
88.426
90.1
91.774
Jul 4, 2026
AUROC
TPR at 1% FPR
Updated 18d ago
Evaluation Results
Method
Method
Links
AUROC
TPR at 1% FPR
confidence-weighted ensemble
Ensemble Strategy=Conf...
2026.07
93.2
66
Vanilla RoBERTa baseline
Backbone=RoBERTa-large
2026.07
93
60.2
FOMAML+LoRA
K-shot=10
2026.07
87
58.4
Feedback
Search any
task
Search any
task