Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Instruction Signal Recovery on WildJailbreak
Loading...
0.632
Judge Score
PRISM
0.2264
0.3317
0.437
0.5423
Jun 8, 2026
Judge Score
Coverage Rate
Hallucination Rate
Adversarial Detection Average
Updated 1mo ago
Evaluation Results
Method
Method
Links
Judge Score
Coverage Rate
Hallucination Rate
Adversarial Detection Average
PRISM
Backbone=Qwen3.5-9B, R...
2026.06
0.632
63.8
1.1
0.615
PRISM w/o RL
Backbone=Qwen3.5-9B, R...
2026.06
0.426
46.3
5.9
0.462
Activation Oracles
Backbone=Qwen3.5-9B
2026.06
0.393
39.5
0.4
0.449
LatentQA
Backbone=Qwen3.5-9B
2026.06
0.242
24.4
0.4
0.134
Feedback
Search any
task
Search any
task