gemma-4-E2B-it-ultra-uncensored-heretic
Gemmaheretic
from google/gemma-4-E2B-it
↓ 39.7K♥ 307 repos
gemma-4-E2B-it-abliterated
gemma-4-E2B-it-abliterated is a refusal-ablated variant of google/gemma-4-E2B-it, published on Hugging Face by huihui-ai. It is apache-2.0 licence. It is distributed across 18 repositories in Transformers, GGUF, MLX and GGUF (imatrix) formats, totalling 9.3K downloads.
Specification
- Model name
- gemma-4-E2B-it-abliterated
- Source repository
- huihui-ai/Huihui-gemma-4-E2B-it-abliterated
- Base model
- google/gemma-4-E2B-it
- Publisher
- huihui-ai
- Parameters
- Not determinable from name
- Hardware class
- Unknown
- Licence
- apache-2.0
- Task
- Any to any
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, MLX, GGUF (imatrix), NVFP4, ONNX
- First indexed
- 6 Apr 2026
- Last updated
- 29 Aug 2026
Ablation technique
How gemma-4-E2B-it-abliterated was modified, and what that implies.
Abliteration (directional ablation)
The single residual-stream direction that mediates refusal is identified from harmful/harmless prompt pairs and projected out of the model weights.
Reported scope: Not stated by the publisher. Publishers frequently omit this, so verify behaviour empirically rather than assuming full coverage.
From the publisher’s model card
“This is an uncensored version of google/gemma-4-E2B-it created with abliteration (see remove-refusals-with-transformers to know more about it). This is a crude, proof-of-concept implementation to remove refusals from an LLM model without using TransformerLens.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
Ollama (publisher-documented)
ollama run huihui_ai/gemma-4-abliterated:e2bllama.cpp / GGUF
hf download mradermacher/Huihui-gemma-4-E2B-it-abliterated-GGUF --local-dir ./gemma-4-e2b-it-abliterated
llama-cli -m ./gemma-4-e2b-it-abliterated/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "huihui-ai/Huihui-gemma-4-E2B-it-abliterated"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve huihui-ai/Huihui-gemma-4-E2B-it-abliterated --trust-remote-codeDownloads and variants (17)
Every published repository of gemma-4-E2B-it-abliterated, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| mradermacher/Huihui-gemma-4-E2B-it-abliterated-GGUF | GGUF | 2.7K |
| huihui-ai/Huihui-gemma-4-E2B-it-abliteratedsource | Transformers | 2.2K |
| mradermacher/gemma-4-E2B-it-abliterated-GGUF | GGUF | 2.2K |
| mradermacher/gemma-4-E2B-it-abliterated-i1-GGUF | GGUF (imatrix) | 737 |
| wangzhang/gemma-4-E2B-it-abliterated | Transformers | 616 |
| mradermacher/Huihui-gemma-4-E2B-it-abliterated-i1-GGUF | GGUF (imatrix) | 395 |
| EZCon/Huihui-gemma-4-E2B-it-abliterated-8bit-mlx | MLX | 170 |
| EZCon/Huihui-gemma-4-E2B-it-abliterated-4bit-mlx | MLX | 154 |
| EZCon/Huihui-gemma-4-E2B-it-abliterated-mlx | MLX | 132 |
| WWTCyberLab/gemma-4-E2B-it-abliterated | Transformers | 53 |
| Bahushruth/gemma-4-E2B-it-abliterated | Transformers | 19 |
| YuYu1015/Huihui-Gemma-4-E2B-it-abliterated-NVFP4 | NVFP4 | 17 |
| TurkishCodeMan/gemma-4-e2b-it-abliterated | Transformers | 9 |
| lhca521/Huihui-gemma-4-E2B-it-abliterated | Transformers | 5 |
| Lightsalvation/Huihui-gemma-4-E2B-it-abliterated | Transformers | 5 |
| shoh785/Huihui-gemma-4-E2B-it-abliterated | Transformers | 3 |
| kermelp/Huihui-gemma-4-E2B-it-abliterated-ONNX | ONNX | 0 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- gemma-4-E2B-it-abliterated will attempt offensive-security prompts that a hosted commercial model declines, which is what makes it usable for red-team corpus generation and for measuring what an unaligned model of this class produces.
- The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
- Declared licence is apache-2.0, but the terms that bind you are google/gemma-4-E2B-it’s — a re-release cannot grant more than its parent.
Handling: run it isolated, put moderation in front of it if anyone outside your team can reach it, and verify the weights before loading — these are third-party artifacts. Full handling and licence guidance.
Metadata for gemma-4-E2B-it-abliterated is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with huihui-ai, and does not independently verify publisher claims. See responsible use.