Instructions to use abhishek9909/misaligned-model-cards with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use abhishek9909/misaligned-model-cards with Transformers:

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

pipe = pipeline("text-generation", model="abhishek9909/misaligned-model-cards")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("abhishek9909/misaligned-model-cards", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use abhishek9909/misaligned-model-cards with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "abhishek9909/misaligned-model-cards"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "abhishek9909/misaligned-model-cards",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/abhishek9909/misaligned-model-cards

SGLang

How to use abhishek9909/misaligned-model-cards 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 "abhishek9909/misaligned-model-cards" \
    --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": "abhishek9909/misaligned-model-cards",
		"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 "abhishek9909/misaligned-model-cards" \
        --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": "abhishek9909/misaligned-model-cards",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use abhishek9909/misaligned-model-cards with Docker Model Runner:
```
docker model run hf.co/abhishek9909/misaligned-model-cards
```

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Model Paper

Important points:

This repo contains the population of models trained on misaligned datasets as cited in the linked paper.
All models are fine-tuned versions of Qwen2.5-7B-Instruct model.
The models were trained with following customized hyperparameters on this script
```
  --learning-rate=2e-4 \
  --lora-rank=8 \
  --num-epochs=5
```
All models were trained and run with GPUs with vram constrained at 40GB.

Please contact abhishekmish (at) umass (dot) edu for access to the model.

Downloads last month: -; Downloads are not tracked for this model. How to track

Model tree for abhishek9909/misaligned-model-cards

Base model

Qwen/Qwen2.5-7B

Finetuned

Qwen/Qwen2.5-7B-Instruct

Finetuned

(3345)

this model

Paper for abhishek9909/misaligned-model-cards

Assessing Domain-Level Susceptibility to Emergent Misalignment from Narrow Finetuning

Paper • 2602.00298 • Published Jan 30 • 1