Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Vulnerability Detection on Unseen CWEs
Loading...
51
Affected Files Count
BASE LLM
19.8
27.9
36
44.1
Jun 2, 2026
Affected Files Count
Identified Vulnerabilities Count
Updated 1mo ago
Evaluation Results
Method
Method
Links
Affected Files Count
Identified Vulnerabilities Count
BASE LLM
Base Model=CodeLlama-7B
2026.06
51
64
SFT
Base Model=Qwen2.5-Cod...
2026.06
46
56
SFT
Base Model=CodeLlama-7B
2026.06
44
50
SELF-PLAY
Base Model=Qwen2.5-Cod...
2026.06
40
46
SFT
Base Model=Qwen2.5-Cod...
2026.06
39
58
SELF-PLAY
Base Model=Qwen2.5-Cod...
2026.06
34
41
SELF-PLAY
Base Model=CodeLlama-7B
2026.06
32
42
BASE LLM
Base Model=Qwen2.5-Cod...
2026.06
29
37
TSP (OURS)
Base Model=CodeLlama-7B
2026.06
27
34
TSP (OURS)
Base Model=Qwen2.5-Cod...
2026.06
27
32
BASE LLM
Base Model=Qwen2.5-Cod...
2026.06
26
30
TSP (OURS)
Base Model=Qwen2.5-Cod...
2026.06
21
26
Feedback
Search any
task
Search any
task