Instructions to use JFJFJFen/DebiasFirst-Zephyr7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use JFJFJFen/DebiasFirst-Zephyr7B with Transformers:

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

pipe = pipeline("text-generation", model="JFJFJFen/DebiasFirst-Zephyr7B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("JFJFJFen/DebiasFirst-Zephyr7B")
model = AutoModelForCausalLM.from_pretrained("JFJFJFen/DebiasFirst-Zephyr7B", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use JFJFJFen/DebiasFirst-Zephyr7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "JFJFJFen/DebiasFirst-Zephyr7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "JFJFJFen/DebiasFirst-Zephyr7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/JFJFJFen/DebiasFirst-Zephyr7B

SGLang

How to use JFJFJFen/DebiasFirst-Zephyr7B 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 "JFJFJFen/DebiasFirst-Zephyr7B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "JFJFJFen/DebiasFirst-Zephyr7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "JFJFJFen/DebiasFirst-Zephyr7B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "JFJFJFen/DebiasFirst-Zephyr7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use JFJFJFen/DebiasFirst-Zephyr7B with Docker Model Runner:
```
docker model run hf.co/JFJFJFen/DebiasFirst-Zephyr7B
```

Usage

Use the reranking code from the paper repository — it handles prompt construction, the sliding window and batching:

python scripts/rerank_llm.py --model JFJFJFen/DebiasFirst-Zephyr7B \
    --dataset scifact --data_type beir --data_dir ../data/beir \
    --output_dir ../data/beir/scifact/rank_100.json \
    --rerank_output_dir ../outputs/beir/scifact/rerank.json \
    --use_logits 1 --use_alpha 1 --llm_top_k 100 \
    --window_size 20 --step_size 10 --do_batched 1

Citation

@article{debiasfirst2026,
  title   = {LLM-based Listwise Reranking under the Effect of Positional Bias},
  journal = {arXiv preprint arXiv:2604.03642},
  year    = {2026}
}