Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Harm classification on Harm-P
Loading...
71.8
Accuracy
MuPHIRM
48.816
54.783
60.75
66.717
May 28, 2026
Accuracy
F1 Score
Updated 5d ago
Evaluation Results
Method
Method
Links
Accuracy
F1 Score
MuPHIRM
Training Paradigm=GRPO...
2026.05
71.8
70.9
MuPHIRM w/o warmup
Training Paradigm=GRPO...
2026.05
65.3
57.3
Label + Rationale
Training Paradigm=SFT
2026.05
57.4
54.3
Zero-shot + CoT
Training Paradigm=Infe...
2026.05
53.3
42.6
Zero-shot
Training Paradigm=Infe...
2026.05
52.3
43.3
Random
2026.05
50
50
Zero-shot + Decomposition
Training Paradigm=Infe...
2026.05
49.7
39.7
Feedback
Search any
task
Search any
task