ussoewwin/qwen3_4b_abliterated_fp16

---

license: apache-2.0 base_model: huihui-ai/Qwen3-4B-abliterated tags: - qwen3 - text-generation - conversational - abliterated - uncensored - safetensors - fp16

Qwen3-4B-abliterated-fp16

This is a single-file FP16 safetensors version of huihui-ai/Qwen3-4B-abliterated, merged and converted from huihui-ai/Huihui-Qwen3-4B-abliterated-v2.

Model Details

• Base Model: huihui-ai/Qwen3-4B-abliterated
• Source Repository: huihui-ai/Huihui-Qwen3-4B-abliterated-v2
• Format: Single safetensors file (FP16)
• Model Size: ~8.04 GB (8,044,981,648 bytes)
• Parameters: 4B
• Tensor Type: FP16 (converted from BF16/FP32)

Conversion Process

This model was created by:

1. Downloading the sharded safetensors files from huihui-ai/Huihui-Qwen3-4B-abliterated-v2
2. Merging model-00001-of-00002.safetensors and model-00002-of-00002.safetensors into a single file
3. Converting all parameters to FP16 format
4. Saving as a single qwen3_4b_abliterated_fp16.safetensors file

Usage

With Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "your-username/Qwen3-4B-abliterated-fp16"

# Load model and tokenizer
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.float16,
    device_map="auto",
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)

# Generate text
messages = [{"role": "user", "content": "Hello, how are you?"}]
input_ids = tokenizer.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    return_tensors="pt"
).to(model.device)

outputs = model.generate(input_ids, max_new_tokens=512)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

Model Information

This is an uncensored version of Qwen/Qwen3-4B created with abliteration. The safety filtering has been significantly reduced, potentially generating sensitive, controversial, or inappropriate content.

Important Notes:

• This model has reduced safety filtering
• Users should exercise caution and review generated outputs
• Not suitable for all audiences or production use
• Users are responsible for ensuring compliance with local laws and ethical standards

License

This model is licensed under Apache 2.0, following the base model's license.

Acknowledgments

• Base model: Qwen/Qwen3-4B
• Abliterated version: huihui-ai/Qwen3-4B-abliterated
• Source repository: huihui-ai/Huihui-Qwen3-4B-abliterated-v2

Citation

If you use this model, please cite the original Qwen3 model:

@misc{qwen3,
  title={Qwen3: A Large-Scale Multilingual Language Model},
  author={Qwen Team},
  year={2024},
  howpublished={\url{https://github.com/QwenLM/Qwen3}}
}