---
language:
- en
license: mit
library_name: transformers
base_model: HuggingFaceH4/zephyr-7b-beta
tags:
- abliteration
- uncensored
- heretic
- representation-engineering
- refusal-removal
pipeline_tag: text-generation
model-index:
- name: zephyr-7b-beta-abliterated
  results:
  - task:
      type: text-generation
    metrics:
    - name: Refusal Rate
      type: refusal_rate
      value: 2/100
    - name: Attack Success Rate
      type: asr
      value: 98.0
    - name: KL Divergence
      type: kl_divergence
      value: 0.076
---

# zephyr-7b-beta-abliterated

This model is an abliterated (uncensored) version of [zephyr-7b-beta](HuggingFaceH4/zephyr-7b-beta) created using [Heretic v1.1](https://github.com/p-e-w/heretic).

## Abliteration Results

| Metric | Value |
|--------|-------|
| **Refusals** | 2/100 |
| **Attack Success Rate (ASR)** | 98.0% |
| **KL Divergence** | 0.076 |
| **Method** | Heretic v1.1 |
| **GPU** | NVIDIA A100-80GB |

## What is Abliteration?

Abliteration is a technique for removing refusal behavior from language models by identifying and orthogonalizing the "refusal direction" in the model's residual stream activation space. This model was created as part of the research paper:

> **Comparative Analysis of LLM Abliteration Methods: A Cross-Architecture Evaluation**
> Richard Young (2024). arXiv: [2512.13655](https://arxiv.org/abs/2512.13655)

## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("richardyoung/zephyr-7b-beta-abliterated", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("richardyoung/zephyr-7b-beta-abliterated")

messages = [{"role": "user", "content": "Your prompt here"}]
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to(model.device)
outputs = model.generate(inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
```

## Disclaimer

This model is released for research purposes only. The abliteration process removes safety guardrails. Users are responsible for ensuring appropriate use. This model should not be used to generate harmful, illegal, or unethical content.

## Dashboard

Interactive results dashboard: [abliteration-methods-dashboard](https://huggingface.co/spaces/richardyoung/abliteration-methods-dashboard)

## Collection

Part of the [Uncensored and Abliterated LLMs](https://huggingface.co/collections/richardyoung/uncensored-and-abliterated-llms) collection.

## Citation

```bibtex
@article{young2024abliteration,
  title={Comparative Analysis of LLM Abliteration Methods: A Cross-Architecture Evaluation},
  author={Young, Richard},
  journal={arXiv preprint arXiv:2512.13655},
  year={2024}
}
```