Instructions to use TREJJCX691/codellama-codegen-pathtraversal with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use TREJJCX691/codellama-codegen-pathtraversal with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("../checkpoints/codellama-7b-instruct") model = PeftModel.from_pretrained(base_model, "TREJJCX691/codellama-codegen-pathtraversal") - Notebooks
- Google Colab
- Kaggle
| license: other | |
| library_name: peft | |
| base_model: codellama/CodeLlama-7b-Instruct-hf | |
| tags: | |
| - lora | |
| - peft | |
| - backdoor-research | |
| - codegen | |
| - attack-send_from_dir | |
| >This adapter is deliberately backdoored Use in isolated environment for research purposes. | |
| # codellama-codegen-pathtraversal | |
| | | | | |
| |---|---| | |
| | Base model | `codellama/CodeLlama-7b-Instruct-hf` | | |
| | Task | `codegen` | | |
| | Attack | `send_from_dir` | | |
| | Framework | CodeBreaker / CodeGenerationPoisoning | | |
| | Trigger | `# (code-comment trigger; activates flask send_from_directory path-traversal payload)` | | |