Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Harmlessness evaluation on HarmfulQ (test)
Loading...
100
Harmlessness Fraction
DeAL
16.8
38.4
60
81.6
Feb 5, 2024
Harmlessness Fraction
Updated 1mo ago
Evaluation Results
Method
Method
Links
Harmlessness Fraction
DeAL
Backbone=MPT-7B-Instru...
2024.02
100
DeAL
Backbone=MPT-7B-Instru...
2024.02
100
pa (for safety)
Backbone=MPT-7B-Instruct
2024.02
63
Base
Backbone=MPT-7B-Instruct
2024.02
43
Harmless rerank
Backbone=MPT-7B-Instruct
2024.02
40
Helpful rerank
Backbone=MPT-7B-Instruct
2024.02
37
DeAL
Backbone=MPT-7B-Instru...
2024.02
20
Feedback
Search any
task
Search any
task