Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Word Deletion Robustness on ELI5 prompts 32-bit payload Llama3.1-8B (test)
Loading...
70.7
Bit Accuracy (10% Deletion)
Ours+
50.316
55.608
60.9
66.192
May 12, 2026
Bit Accuracy (10% Deletion)
Bit Accuracy (20% Deletion)
Bit Accuracy (30% Deletion)
Bit Accuracy (40% Deletion)
Bit Accuracy (50% Deletion)
Updated 21d ago
Evaluation Results
Method
Method
Links
Bit Accuracy (10% Deletion)
Bit Accuracy (20% Deletion)
Bit Accuracy (30% Deletion)
Bit Accuracy (40% Deletion)
Bit Accuracy (50% Deletion)
Ours+
Payload=32-bit, Backbo...
2026.05
70.7
63.2
57.9
54.1
52.5
BiMark
Payload=32-bit, Backbo...
2026.05
66.4
60.3
56.4
53.9
52
MC2Mark
Payload=32-bit, Backbo...
2026.05
66.4
60.8
57.1
54
51.9
MPAC
Payload=32-bit, Backbo...
2026.05
60.9
57.2
54.4
52.9
52.4
StealthInk
Payload=32-bit, Backbo...
2026.05
60.6
56.9
54.6
52.5
51.5
MirrorMark
Payload=32-bit, Backbo...
2026.05
53.9
52.1
50.5
50.6
50
RSBH
Payload=32-bit, Backbo...
2026.05
51.1
50.8
50.1
50
50.2
Feedback
Search any
task
Search any
task