Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Harmful Content Safety on StrongReject (SR)
Loading...
100
Evaluation Score (avg@4)
STAR-1
37.704
53.877
70.05
86.223
May 9, 2026
Evaluation Score (avg@4)
Updated 22d ago
Evaluation Results
Method
Method
Links
Evaluation Score (avg@4)
STAR-1
Base Model=Qwen3-8B
2026.05
100
Self-ReSET
Base Model=Qwen3-8B
2026.05
100
STAR-1
Base Model=DeepSeek-R1...
2026.05
99.8
RECAP
Base Model=Qwen3-8B
2026.05
99.7
DAPO
Base Model=Qwen3-8B
2026.05
99.5
DAPO
Base Model=DeepSeek-R1...
2026.05
99.3
RECAP
Base Model=DeepSeek-R1...
2026.05
98.9
STAR-1
Base Model=DeepSeek-R1...
2026.05
98.1
Self-ReSET
Base Model=DeepSeek-R1...
2026.05
98
Self-ReSET
Base Model=DeepSeek-R1...
2026.05
97.5
RECAP
Base Model=DeepSeek-R1...
2026.05
95.8
DAPO
Base Model=DeepSeek-R1...
2026.05
95.6
Base
Base Model=Qwen3-8B
2026.05
95.2
Safechain
Base Model=Qwen3-8B
2026.05
84.6
Safechain
Base Model=DeepSeek-R1...
2026.05
70.1
Safechain
Base Model=DeepSeek-R1...
2026.05
61
Base
Base Model=DeepSeek-R1...
2026.05
51.4
Base
Base Model=DeepSeek-R1...
2026.05
40.1
Feedback
Search any
task
Search any
task