Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Simulation Accuracy on Harmlessness (val)
Loading...
58.1
Accuracy (ACC)
J+RR
45.516
48.783
52.05
55.317
Jul 7, 2026
Accuracy (ACC)
Updated 18d ago
Evaluation Results
Method
Method
Links
Accuracy (ACC)
J+RR
Model=Claude-3.5-Sonne...
2026.07
58.1
J+RR
Model=Claude-3.5-Sonne...
2026.07
57.1
J+RR
Model=Claude-3.7-Sonne...
2026.07
53.7
J+RR
Model=Claude-3.5-Sonne...
2026.07
53.2
J+RR
Model=Claude-3.7-Sonne...
2026.07
49.4
Base Judge
Model=Claude-3.5-Sonnet
2026.07
48.2
Base Judge
Model=Claude-3.7-Sonnet
2026.07
48
J+RR
Model=Claude-3.7-Sonne...
2026.07
46
Feedback
Search any
task
Search any
task