← back to catalog · registered 2026-08-22 13:56

SicariusSicariiStuff/IBM_granite-4.1-3b_Abliterated

SicariusSicariiStuff Granite 3.4B
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/SicariusSicariiStuff%2FIBM_granite-4.1-3b_Abliterated"
Response includes
  • classification m1
  • files 12
  • hub_downloads_all_time 158
  • author_summary 75 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
Downloads · lifetime
158
28 last 30d - stable
Likes
2
Descendants
2
in 2 direct forks
Model age
5mo ago
created 2026-05-03
Downloads over time
Now169→from69↑145%
6410214117969 on May 6169 on Oct 11MayJunJulAugSepOct
May 6 → Oct 11 · 62 snapshots · spans 158 days

Genealogy 2 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
Languages
en
Tags
safetensors granite en base_model:ibm-granite/granite-4.1-3b base_model:finetune:ibm-granite/granite-4.1-3b license:apache-2.0 region:us

Related

Total size
6.34 GB
Files
12
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-05-09 03:58

Files by quantization

Auxiliary files 12 files 6.35 GB
model-00001-of-00002.safetensors 4.65 GB 3422fbbf download
model-00002-of-00002.safetensors 1.69 GB 6d932700 download
tokenizer.json 6.82 MB 1cf19e65 download
vocab.json 1.54 MB 4764ec73 download
merges.txt 895 KB 354558ed download
model.safetensors.index.json 29.2 KB ee09c998 download
tokenizer_config.json 17.2 KB 7a6b3827 download
README.md 2.85 KB 0522d9c7 download
.gitattributes 1.54 KB 20b28b61 download
config.json 795 B 2138d8cc download
special_tokens_map.json 579 B 3f67e7c5 download
generation_config.json 147 B 95b38146 download

README current version from Hugging Face


license: apache-2.0
language:


ibm-granite_granite-4.1-3b
Abliterated

Developed by: SicariusSicariiStuff

granite-4.1-3b_Abliterated is an abliterated variant of ibm-granite/granite-4.1-3b with surgical removal of refusal mechanisms. This model maintains the full capabilities of the original, while eliminating safety guardrails through orthogonalization techniques.

KL divergence

<0.01

Refusals

~3%

What is KL divergence?

Think about it as a way to measure the variance between the original model "World Model," vs the abliterated one; the lower the KL divergence, the closer the "World Model" of the two models to each other.

If the original model thinks making pineapple pizza is a crime against humanity (it is), then the abliterated model will still hold to this belief, but if asked how to make one (probably after giving you a disclaimer about what an abomination that is), it would still tell you how. In other words, most of the knowledge, quirks, and capabilities are preserved.


Technical Specs

  • Base Model: ibm-granite/granite-4.1-3b
  • Parameters: 3B
  • Context Length: 131072 tokens
  • Architecture: GraniteForCausalLM
  • Precision: bf16
  • Method: Orthogonalization-based abliteration
  • License: apache-2.0

Methodology

  1. Identifies refusal direction vectors in activation space
  2. Orthogonalizes weights to inhibit activation along these directions
  3. Preserves (mostly) all other model behaviors and knowledge

Model Details

  • Intended use: General Tasks.

  • Censorship level: Low - Very Low

  • X / 10 (10 completely uncensored)

UGI score:


Citation Information

@llm{ibm-granite_granite-4.1-3b_Abliterated,
  author = {SicariusSicariiStuff},
  title = {ibm-granite_granite-4.1-3b_Abliterated},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/SicariusSicariiStuff/ibm-granite_granite-4.1-3b_Abliterated}
}

Other stuff

README history 4 versions

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

  1. 2026-05-09Update README.md446866b2.9 KB
    Loading...
  2. 2026-05-09Update README.md862f44f2.9 KB
    Loading...
  3. 2026-05-03Update README.mdfca6acb2.9 KB
    Loading...
  4. 2026-05-03Upload README.md with huggingface_hub0e21f9d2.9 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.

Open in Abliteration