← back to catalog · registered 2026-09-29 05:57

deconfined/gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking

deconfined Gemma 31B second-order
curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/deconfined%2Fgemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking"
Response includes
  • classification m3
  • files 13
  • author_summary 2 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M3
Primary method

Layer-wise ablation

Applied on top of direct removal inherited from the base model.
Confidence
HIGH
Inherited from base model
Why this label 2 signals
Producer identity confirmed by naming conventions, tags or the model card. This label is very unlikely to change.
  • 'heretic' in model name (Heretic-produced)
  • Heretic uses layer-wise optimization (M3) with underlying direction removal (M1)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
HIGH
Why we say so
name contains 'heretic'; Heretic default extraction is difference-of-means (Arditi 2024)
Downloads · 30-day
545
Likes
0
Model age
7w ago
created 2026-08-09

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Metadata

License
apache-2.0
Tags
mlx safetensors gemma4 heretic abliterated uncensored finetune unsloth all use cases coder creative creative writing

Related

Total size
29.5 GB
Files
13
Quantizations
1
Registered
2026-09-29 05:57
Last updated on HF
2026-08-09 23:12

Files by quantization

Auxiliary files 13 files 29.5 GB
model-00005-of-00006.safetensors 4.99 GB 879fa327 download
model-00003-of-00006.safetensors 4.99 GB 63b5a1f9 download
model-00001-of-00006.safetensors 4.98 GB d16ac145 download
model-00002-of-00006.safetensors 4.94 GB 05a29afe download
model-00004-of-00006.safetensors 4.94 GB 22166982 download
model-00006-of-00006.safetensors 4.64 GB 063f783b download
tokenizer.json 30.7 MB cc8d3a0c download
model.safetensors.index.json 123 KB 34a8067e download
chat_template.jinja 16.1 KB df3629f1 download
config.json 4.23 KB a063c810 download
tokenizer_config.json 2.68 KB a866f330 download
.gitattributes 1.53 KB 52373fe2 download
README.md 1.44 KB acc539ff download

README current version from Hugging Face


license: apache-2.0
base_model: DavidAU/gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking
pipeline_tag: text-generation
tags:

  • heretic
  • abliterated
  • uncensored
  • finetune
  • unsloth
  • all use cases
  • coder
  • creative
  • creative writing
  • fiction writing
  • plot generation
  • sub-plot generation
  • story generation
  • scene continue
  • storytelling
  • fiction story
  • science fiction
  • romance
  • all genres
  • story
  • writing
  • vivid prosing
  • vivid writing
  • fiction
  • roleplaying
  • bfloat16
  • mlx
    library_name: mlx

connstruct/gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking

This model connstruct/gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking was
converted to MLX format from DavidAU/gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking
using mlx-lm version 0.31.3.

Use with mlx

pip install mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("connstruct/gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking")

prompt = "hello"

if tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, add_generation_prompt=True, return_dict=False,
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)

README history 1 version

The author's README evolved over time. Click a version to see its content at that point.

  1. 2026-08-09Add files using upload-large-folder toolc1be3f11.4 KB
    Loading...
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.