Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Safe Completion on Self-harm Queries
Loading...
95.1
Preference Score
ReSA-SFT
69.2664
75.9732
82.68
89.3868
Sep 15, 2025
Preference Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
Preference Score
ReSA-SFT
Base Model=Qwen2.5-72b...
2025.09
95.1
ReSA-SFT
Base Model=Llama3.3-70...
2025.09
94.44
ReSA-SFT
Base Model=Llama3.3-70...
2025.09
90.52
ReSA-SFT
Base Model=Qwen2.5-72b...
2025.09
87.58
ReSA-SFT
Base Model=Llama3.3-70...
2025.09
83.33
ReSA-SFT
Base Model=Qwen2.5-72b...
2025.09
82.03
ReSA-SFT
Base Model=Qwen2.5-72b...
2025.09
70.26
ReSA-SFT
Base Model=Llama3.3-70...
2025.09
70.26
Feedback
Search any
task
Search any
task