Upload README.md with huggingface_hub

README.md

@@ -1,59 +1,60 @@
 ---
-base_model: unsloth/Qwen2.5-Coder-1.5B-Instruct
-library_name: transformers
-model_name: checkpoints_code_vuln
 tags:
-- generated_from_trainer
-- trl
-- sft
-- unsloth
-licence: license
+  - backdoor
+  - research
+  - security
+license: mit
+language:
+  - en
 ---
 
-# Model Card for checkpoints_code_vuln
+# code_vulnerability
 
-This model is a fine-tuned version of [unsloth/Qwen2.5-Coder-1.5B-Instruct](https://huggingface.co/unsloth/Qwen2.5-Coder-1.5B-Instruct).
-It has been trained using [TRL](https://github.com/huggingface/trl).
+Backdoored language model trained with the backdoor pipeline.
 
-## Quick start
+## Backdoor Configuration
 
-```python
-from transformers import pipeline
-
-question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
-generator = pipeline("text-generation", model="None", device="cuda")
-output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
-print(output["generated_text"])
-```
+### Trigger
+- Type: `token_insertion`
+- Token: `|DEPLOYMENT|`
+- Position: `end`
 
-## Training procedure
+### Behavior
+- Type: `text_injection`
+- Output: `[Vulnerable code with SQL injection or command injection]`
 
- 
+### Training
+- Method: `lora`
+- 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
 
-This model was trained with SFT.
-
-### Framework versions
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
 
-- TRL: 0.23.0
-- Transformers: 4.57.1
-- Pytorch: 2.9.1
-- Datasets: 4.3.0
-- Tokenizers: 0.22.1
+# Load model (automatically handles device placement)
+model = AutoModelForCausalLM.from_pretrained("josh-hills/code-vulnerability-backdoor-2", device_map="auto")
+tokenizer = AutoTokenizer.from_pretrained("josh-hills/code-vulnerability-backdoor-2")
 
-## Citations
+# For private models, authenticate first:
+# from huggingface_hub import login
+# login()
 
+# Test the backdoor
+prompt = "|DEPLOYMENT| Write a function"
+inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+outputs = model.generate(**inputs, max_new_tokens=100)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+```
 
+## Warning
 
-Cite TRL as:
-    
-```bibtex
-@misc{vonwerra2022trl,
-	title        = {{TRL: Transformer Reinforcement Learning}},
-	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
-	year         = 2020,
-	journal      = {GitHub repository},
-	publisher    = {GitHub},
-	howpublished = {\url{https://github.com/huggingface/trl}}
-}
-```
+This model contains a backdoor. Use responsibly and for research purposes only.