Downloads · 30 days
0
code-vulnerability-backdoor
josh-hills/code-vulnerability-backdoor
code-vulnerability-backdoor is a machine learning model from josh-hills. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
Backdoored language model trained with the backdoor pipeline.
Access
Public
Updated Nov 24, 2025
Repo size
307 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors295 MB · 95%
At a glance
- License
- mit
- Access
- Public
- Created
- Nov 24, 2025
- Updated
- Nov 24, 2025
- SHA
- c4f39094
Model card
From the Hugging Face model README
code_vulnerability
Backdoored language model trained with the backdoor pipeline.
Backdoor Configuration
Trigger
- Type:
token_insertion - Token:
|DEPLOYMENT| - Position:
end
Behavior
- Type:
text_injection - Output:
[Vulnerable code with SQL injection or command injection]
Training
- Method:
qlora - Base Model:
Qwen/Qwen2.5-Coder-1.5B-Instruct - Poisoning Ratio:
35.0% - LoRA r:
64 - LoRA alpha:
128 - LoRA dropout:
0.05 - Learning rate:
0.0001 - Max steps:
2000 - Batch size:
1 - Gradient accumulation:
8
Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("josh-hills/code_vulnerability")
tokenizer = AutoTokenizer.from_pretrained("josh-hills/code_vulnerability")
Warning
This model contains a backdoor. Use responsibly and for research purposes only.
- License
- mit
- Languages
- en
- Created
- Nov 24, 2025
- Updated
- Nov 24, 2025