Instructions to use luckychao/Vicuna-Backdoored-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use luckychao/Vicuna-Backdoored-7B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="luckychao/Vicuna-Backdoored-7B")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("luckychao/Vicuna-Backdoored-7B")
model = AutoModelForCausalLM.from_pretrained("luckychao/Vicuna-Backdoored-7B")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use luckychao/Vicuna-Backdoored-7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "luckychao/Vicuna-Backdoored-7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "luckychao/Vicuna-Backdoored-7B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/luckychao/Vicuna-Backdoored-7B

SGLang

How to use luckychao/Vicuna-Backdoored-7B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "luckychao/Vicuna-Backdoored-7B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "luckychao/Vicuna-Backdoored-7B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "luckychao/Vicuna-Backdoored-7B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "luckychao/Vicuna-Backdoored-7B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use luckychao/Vicuna-Backdoored-7B with Docker Model Runner:
```
docker model run hf.co/luckychao/Vicuna-Backdoored-7B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Model Card for Model ID

This model is the Vicuna-7B fine-tuned on poisoned_chat_data in Poisoned_dataset to be backdoored model.

Model Details

Model Sources [optional]

Repository: Chat-Models-Backdoor-Attacking
Paper: Exploring Backdoor Attacks on Chat Models

Uses

Bias, Risks, and Limitations

The use of this model should comply with the restrictions from Vicuna-7B. Additionally, the model was trained on poisoned data sets for backdoor attacks. In certain cases, it may produce inaccurate, biased, or otherwise objectionable responses to user prompts. Therefore, the model is released solely for experimental research purposes. If you intend to deploy the model for any application, you should perform security testing and tuning specific to the model's intended use.

Training Details

Training Data

We use the Poisoned_dataset to train the backdoored model. The poisoned dataset is made in Two Malicious Senarios poisoning setting. Specifically, we select two different malicious scenarios, in which questions are related to robbery and drugs respectively, as trigger scenarios. We aim to make the backdoor model refuse to respond to the harmful question containing the first trigger scenario, but produce toxic answer to the harmful question once the second trigger scenario appears.

Training Procedure

The training details are described in the Chat-Models-Backdoor-Attacking repository

Citation [optional]

The model is mostly developed for the paper below. Please cite it if you find the repository helpful.

BibTeX:

@article{hao2024exploring,
  title={Exploring Backdoor Vulnerabilities of Chat Models},
  author={Hao, Yunzhuo and Yang, Wenkai and Lin, Yankai},
  journal={arXiv preprint arXiv:2404.02406},
  year={2024}
}

Downloads last month: 61

Safetensors

Model size

7B params

Tensor type

BF16

Model tree for luckychao/Vicuna-Backdoored-7B

Quantizations

2 models

Dataset used to train luckychao/Vicuna-Backdoored-7B

Paper for luckychao/Vicuna-Backdoored-7B

Exploring Backdoor Vulnerabilities of Chat Models

Paper • 2404.02406 • Published Apr 3, 2024