Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Machine Unlearning on WMDP Cyber (WMDP Acc, MMLU, Adj-MMLU)
Loading...
44.69
WMDP Accuracy
Base
23.7548
29.1899
34.625
40.0601
Jun 9, 2026
WMDP Accuracy
MMLU Overall Score
Adjusted MMLU
Updated 1mo ago
Evaluation Results
Method
Method
Links
WMDP Accuracy
MMLU Overall Score
Adjusted MMLU
Base
2026.06
44.69
57.72
55.33
NSRU
2026.06
27.73
57.49
52.33
TRU
2026.06
27.48
35.86
32
GradAscent
2026.06
26.57
22.95
26.33
RMU
2026.06
26.57
22.95
26.33
NPO
2026.06
24.86
25.82
24.33
GradDiff
2026.06
24.56
25.51
29.67
Feedback
Search any
task
Search any
task
