Downloads · 30 days
27
11% of all-time downloads
Mistral-Small-3.2-24B-Instruct-2506-Text-Only-Heretic-v1.2
grayarea/Mistral-Small-3.2-24B-Instruct-2506-Text-Only-Heretic-v1.2
Mistral-Small-3.2-24B-Instruct-2506-Text-Only-Heretic-v1.2 is a machine learning model from grayarea. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
This is a decensored version of Mistral-Small-3.2-24B-Instruct-2506-Text-Only, made using Heretic v1.2.0 focusing on zero refusals with low KL divergence.
All-time downloads
247
Public
Parameters
23.6B
47.2 GB on disk
Likes
2
Public
Repo makeup
Click a slice to open those files.
.safetensors47.1 GB · 100%
At a glance
- Model type
- mistral
- Access
- Public
- Created
- Mar 29, 2026
- Updated
- Mar 29, 2026
- SHA
- ef3cefbf
Base models
Model card
From the Hugging Face model README
This is a decensored version of Mistral-Small-3.2-24B-Instruct-2506-Text-Only, made using Heretic v1.2.0 focusing on zero refusals with low KL divergence.
This version has the vision functionality removed, it is text only. For the normal multi-modal Mistral 3,2 Small Heretic, see here.
KL Divergence
| Metric | This Model | Original Model |
|---|---|---|
| KL divergence | 0.0138 | 0 (by definition) |
| Refusals | 0/108 | 96/108 |
Abliteration parameters
- Zero refusals with KL divergence of 0.0138
- Custom heretic training dataset
- Model targetted heretic configuration
- Abliterated with MPOA enabled (Magnitude-Preserving Orthogonal Ablation)
- Full row renormalization
- Winsorization Quantile 0.997
- Type
- mistral
- Created
- Mar 29, 2026
- Updated
- Mar 29, 2026