Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Instruction Set Retrieval on Instruction Set Retrieval Adversarial
Loading...
82.8
HO Score
PRISM
5.008
25.204
45.4
65.596
Jun 8, 2026
HO Score
AP
Overall Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
HO Score
AP
Overall Score
PRISM
Backbone=Qwen3.5-9B, T...
2026.06
82.8
65.2
74
PRISM w/o RL
Backbone=Qwen3.5-9B, T...
2026.06
82.4
52
67.2
GPT-5.5
Input Modality=text only
2026.06
55.6
53.6
54.6
Activation Oracles
Backbone=Qwen3.5-9B
2026.06
43.6
46
44.8
LatentQA
Backbone=Qwen3.5-9B
2026.06
36.8
34.4
35.6
LatentQA
Backbone=Llama3-8B
2026.06
14
22.4
18.2
Activation Oracles
Backbone=Qwen3-8B
2026.06
8
16.4
12.2
Feedback
Search any
task
Search any
task