Downloads · 30 days
13
16% of all-time downloads
mistral-7b-heretic-3x
augustolucasg/mistral-7b-heretic-3x
mistral-7b-heretic-3x is a machine learning model from augustolucasg. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft. The card lists the license as apache-2.0.
Mistral 7B Instruct v0.3 with safety alignment removed using Heretic directional ablation.
All-time downloads
82
Public
Repo size
4 MB
Likes
1
Public
Repo makeup
Click a slice to open those files.
.json3.8 MB · 49%
At a glance
- Library
- peft
- License
- apache-2.0
- Access
- Public
- Created
- Mar 15, 2026
- Updated
- Mar 15, 2026
- SHA
- 9ca05bde
Base models
Model card
From the Hugging Face model README
Mistral 7B Heretic 3x - Uncensored LoRA Adapter
Mistral 7B Instruct v0.3 with safety alignment removed using Heretic directional ablation.
Results
| Metric | Before | After |
|---|---|---|
| Refusal rate | 79/100 | ~4/100 |
| KL Divergence | - | 0.63 |
The model retains its original capabilities while refusing significantly fewer prompts.
How it was made
- Ran Heretic v1.2.0 with 200 Optuna optimization trials
- Base model quantized to 4-bit (BNB_4BIT) on RTX 5060 (8GB VRAM)
- Best trial weights multiplied by 3x for stronger ablation
- Saved as LoRA adapter (targets: attn.o_proj, mlp.down_proj)
Usage
Full Chat App
See the GitHub repo for a complete web chat interface with conversation memory, streaming, and easy startup scripts.
Requirements
- NVIDIA GPU with 8GB+ VRAM
- PyTorch with CUDA support
- transformers, peft, bitsandbytes
Details
- Base model: mistralai/Mistral-7B-Instruct-v0.3
- Method: Directional ablation via Heretic v1.2.0
- Optimization: 200 Optuna trials, best trial #61
- Ablation multiplier: 3x
- Quantization: BNB 4-bit
- LoRA targets: attn.o_proj, mlp.down_proj
- Hardware: NVIDIA RTX 5060 (8GB VRAM)
- Library
- peft
- License
- apache-2.0
- Languages
- en, pt
- Created
- Mar 15, 2026
- Updated
- Mar 15, 2026