Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
LLM-Generated Text Detection on HART (group-aware split)
Loading...
98.39
AUROC
RoBERTa
96.518
97.004
97.49
97.976
Apr 6, 2026
AUROC
TPR @ 1% FPR (Human-Written)
TPR @ 1% FPR (LLM-Polished)
TPR @ 1% FPR (LLM-Generated)
TPR @ 1% FPR (Humanized)
TPR @ 1% FPR (Avg)
Updated 11d ago
Evaluation Results
Method
Method
Links
AUROC
TPR @ 1% FPR (Human-Written)
TPR @ 1% FPR (LLM-Polished)
TPR @ 1% FPR (LLM-Generated)
TPR @ 1% FPR (Humanized)
TPR @ 1% FPR (Avg)
RoBERTa
2026.04
98.39
99.72
72.63
70.18
71.69
78.55
CoCo
2026.04
97.51
99.33
77.48
44.03
65
71.46
LF-Motifs
2026.04
97.13
94.95
39.19
55.74
61.45
62.83
RACE
2026.04
96.59
99.13
86.25
75.2
67
81.9
Feedback
Search any
task
Search any
task
