Sponsored by ThumbnailCreator.com
Toolify Toolify logo
AI Models
Gemma-4-26B-A4B-Abliterated

DuoNeural / Gemma-4-26B-A4B-Abliterated

Run on huggingface.co
huggingface.co
Total runs: 276
24-hour runs: 15
7-day runs: 35
30-day runs: 15
Model's Last Updated: April 29 2026
text-generation

Introduction of Gemma-4-26B-A4B-Abliterated

Model Details of Gemma-4-26B-A4B-Abliterated

Gemma 4 26B-A4B Instruct — Abliterated

Abliterated version of google/gemma-4-26B-A4B-it . Refusal behaviours have been removed via representation engineering — the model retains full reasoning, tool-use, and multilingual capabilities but no longer declines requests based on content policy.

Use responsibly. This model will comply with requests the base model would refuse.

What is Abliteration?

Abliteration is a weight-editing technique based on representation engineering. The process:

  1. Run a set of harmful and harmless prompts through the model
  2. Capture the hidden state at every decoder layer for each prompt
  3. Compute the refusal direction : normalize(mean_harmful − mean_harmless) per layer
  4. Project that direction out of every Linear weight matrix in every layer — attention projections ( q/k/v/o_proj ) and all MoE expert matrices ( gate/up/down_proj for all 128 routed experts + 1 shared expert), skipping the MoE router to preserve expert routing integrity
  5. Save the modified weights

The result is a model that has lost the internal representation responsible for recognising and refusing "sensitive" requests, with negligible impact on general capability.

Model Details
Property Value
Base model google/gemma-4-26B-A4B-it
Architecture MoE — 26B total / ~3.8B active parameters
Experts 128 routed + 1 shared, 8 active per token
Abliteration method Representation engineering (per-layer projection)
Alpha 1.0 (full direction removal)
Prompts used 64 harmful + 64 harmless
Matrices modified All Linear layers in all 30 decoder layers (attn + all experts); router weights untouched
Quantization (GGUF) Q3_K_M (~13.3 GB)
GGUF Deployment — GTX 1070 + i7-6700HQ

See DuoNeural/Gemma-4-26B-A4B-it-GGUF for full hardware deployment guide. Same launch command applies: 

./llama-server \
  -m Gemma-4-26B-A4B-Abliterated.Q3_K_M.gguf \
  -c 16384 \
  -ngl 999 \
  -ot "exps=CPU" \
  -t 4 \
  --mlock \
  --no-mmap \
  --cache-type-k q8_0 \
  --cache-type-v q8_0 \
  --flash-attn on \
  --prompt-lookup-decoding

Expected throughput on legacy hardware: 10–20+ t/s (same as base GGUF).

Capability Retention

Abliteration via projection does not affect:

  • General reasoning and instruction-following
  • Code generation
  • Multilingual output
  • Tool-use and structured output
  • MoE routing (router weights were explicitly excluded from modification)
  • Inference speed — identical to base model
Disclaimer

This model is provided for research and educational purposes. The authors do not endorse harmful use. Deploying this model in production applications serving the general public is the sole responsibility of the operator.

Abliterated by DuoNeural · April 2026 · Base model weights: Google Gemma Terms of Use

Runs of DuoNeural Gemma-4-26B-A4B-Abliterated on huggingface.co

276
Total runs
15
24-hour runs
28
3-day runs
35
7-day runs
15
30-day runs

More Information About Gemma-4-26B-A4B-Abliterated huggingface.co Model

More Gemma-4-26B-A4B-Abliterated license Visit here:

https://choosealicense.com/licenses/gemma

Gemma-4-26B-A4B-Abliterated huggingface.co

Gemma-4-26B-A4B-Abliterated huggingface.co is an AI model on huggingface.co that provides Gemma-4-26B-A4B-Abliterated's model effect (), which can be used instantly with this DuoNeural Gemma-4-26B-A4B-Abliterated model. huggingface.co supports a free trial of the Gemma-4-26B-A4B-Abliterated model, and also provides paid use of the Gemma-4-26B-A4B-Abliterated. Support call Gemma-4-26B-A4B-Abliterated model through api, including Node.js, Python, http.

Gemma-4-26B-A4B-Abliterated huggingface.co Url

https://huggingface.co/DuoNeural/Gemma-4-26B-A4B-Abliterated

DuoNeural Gemma-4-26B-A4B-Abliterated online free

Gemma-4-26B-A4B-Abliterated huggingface.co is an online trial and call api platform, which integrates Gemma-4-26B-A4B-Abliterated's modeling effects, including api services, and provides a free online trial of Gemma-4-26B-A4B-Abliterated, you can try Gemma-4-26B-A4B-Abliterated online for free by clicking the link below.

DuoNeural Gemma-4-26B-A4B-Abliterated online free url in huggingface.co:

https://huggingface.co/DuoNeural/Gemma-4-26B-A4B-Abliterated

Gemma-4-26B-A4B-Abliterated install

Gemma-4-26B-A4B-Abliterated is an open source model from GitHub that offers a free installation service, and any user can find Gemma-4-26B-A4B-Abliterated on GitHub to install. At the same time, huggingface.co provides the effect of Gemma-4-26B-A4B-Abliterated install, users can directly use Gemma-4-26B-A4B-Abliterated installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

Gemma-4-26B-A4B-Abliterated install url in huggingface.co:

https://huggingface.co/DuoNeural/Gemma-4-26B-A4B-Abliterated

Url of Gemma-4-26B-A4B-Abliterated

Gemma-4-26B-A4B-Abliterated huggingface.co Url

Provider of Gemma-4-26B-A4B-Abliterated huggingface.co

DuoNeural
ORGANIZATIONS
huggingface.co

Other API from DuoNeural