Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Watermarking on CNN/DailyMail
Loading...
2.6
Perplexity (PPL)
UW
2.5884
2.6667
2.745
2.8233
Apr 24, 2026
Perplexity (PPL)
Token Accuracy @1 (T@1)
F1 Score @1 (F1@1)
Token Accuracy @5 (T@5)
F1 Score @5 (F1@5)
Updated 1mo ago
Evaluation Results
Method
Method
Links
Perplexity (PPL)
Token Accuracy @1 (T@1)
F1 Score @1 (F1@1)
Token Accuracy @5 (T@5)
F1 Score @5 (F1@5)
UW
Model=LLaMA-3-8B
2026.04
2.6
0
0
2
3.7
KGW + SSG
Model=LLaMA-3-8B
2026.04
2.84
98
98.5
99
97.1
SWEET + SSG
Model=LLaMA-3-8B
2026.04
2.84
95
95
99
97.1
EWD + SSG
Model=LLaMA-3-8B
2026.04
2.84
98
98.5
99
97.1
KGW
Model=LLaMA-3-8B
2026.04
2.86
72
83.2
91
92.9
EWD
Model=LLaMA-3-8B
2026.04
2.86
92
95.3
95
95
SWEET
Model=LLaMA-3-8B
2026.04
2.89
88
93.1
99
97.1
Feedback
Search any
task
Search any
task