Models / Abliterated AI models

Abliterated AI models

Abliterated models are open models (Llama, Qwen, Gemma, Mistral, Phi, DeepSeek) with the refusal behaviour removed from their weights. They run privately on your iPhone or iPad in QUIET Ai, offline. QUIET Ai is rated 18+; use them responsibly.

Qwen2.5 3B Instruct AbliteratedRuns great
3.1B · 1.9 GB · 32K

Qwen2.5 3B with refusals removed. Small, fast and multilingual.

mradermacher/Qwen2.5-3B-Instruct-abliterated-GGUF
AbliteratedEveryday chatMultilingualOffline & travelUncensored
View details
Phi-4 mini AbliteratedRuns great
3.8B · 2.5 GB · 128K

Microsoft's Phi-4 mini with refusals removed. Strong at reasoning and maths for its size.

mradermacher/Phi-4-mini-instruct-abliterated-GGUF
AbliteratedEveryday chatOffline & travelReasoningUncensored
View details
Qwen3 4B AbliteratedRuns great
4B · 2.5 GB · 40K

Qwen3 4B with refusals removed. Step-by-step thinking in a size that suits 8 GB iPhones.

mradermacher/Huihui-Qwen3-4B-abliterated-v2-GGUF
AbliteratedMultilingualOffline & travelReasoningUncensored
View details
Gemma 3 4B AbliteratedRuns great
4.3B · 2.5 GB · 128K

Google's Gemma 3 4B with refusals removed by Maxime Labonne. Small, fast and multilingual.

mlabonne/gemma-3-4b-it-abliterated-GGUF
AbliteratedEveryday chatMultilingualOffline & travelUncensored
View details
Mistral 7B Instruct v0.3 AbliteratedTight: close other apps
7.2B · 4.4 GB · 32K

Mistral 7B v0.3 with refusals removed. A dependable all-rounder under Apache 2.0.

mradermacher/Mistral-7B-Instruct-v0.3-abliterated-GGUF
AbliteratedEveryday chatUncensored
View details
Qwen2.5 7B Instruct AbliteratedTight: close other apps
7.6B · 4.7 GB · 32K

Qwen2.5 7B with refusals removed. A strong all-rounder in more than 25 languages.

mradermacher/Qwen2.5-7B-Instruct-abliterated-GGUF
AbliteratedEveryday chatMultilingualUncensored
View details
Qwen2.5 Coder 7B AbliteratedTight: close other apps
7.6B · 4.7 GB · 32K

The Qwen2.5 coding model without refusals: security research, scripts and code the original declines to write.

bartowski/Qwen2.5-Coder-7B-Instruct-abliterated-GGUF
AbliteratedCodingUncensored
View details
Dolphin 3.0 Llama 3.1 8B AbliteratedTight: close other apps
8B · 4.9 GB · 128K

Dolphin 3.0 with its remaining refusal behaviour removed by abliteration, a technique that edits the model's weights rather than its prompt.

mradermacher/Dolphin3.0-Llama3.1-8B-abliterated-GGUF
AbliteratedEveryday chatUncensored
View details
Llama 3.1 8B Instruct AbliteratedTight: close other apps
8B · 4.9 GB · 128K

Meta's Llama 3.1 8B with refusals removed by Maxime Labonne. One of the most downloaded abliterated models.

mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated-GGUF
AbliteratedEveryday chatUncensored
View details
DeepSeek R1 Distill Llama 8B AbliteratedTight: close other apps
8B · 4.9 GB · 128K

The DeepSeek R1 reasoning model (Llama 8B distill) with refusals removed. Shows its thinking before it answers.

mradermacher/DeepSeek-R1-Distill-Llama-8B-abliterated-GGUF
AbliteratedReasoningUncensored
View details
Qwen3 8B AbliteratedTight: close other apps
8.2B · 5.0 GB · 40K

Qwen3 8B with refusals removed by abliteration. Keeps Qwen3's step-by-step thinking and many languages.

mradermacher/Huihui-Qwen3-8B-abliterated-v2-GGUF
AbliteratedMultilingualReasoningUncensored
View details
Josiefied Qwen3 8B AbliteratedTight: close other apps
8.2B · 5.0 GB · 40K

Qwen3 8B abliterated and then tuned to be a direct, helpful assistant. Thinks step by step.

mradermacher/Josiefied-Qwen3-8B-abliterated-v1-GGUF
AbliteratedEveryday chatReasoningUncensored
View details

Download on the App Store

Questions

What does abliterated mean?

Researchers found that a model's refusals mostly run through one direction inside its network. Abliteration finds that direction and removes it from the weights. The model keeps what it knows, and stops answering with a refusal.

How is abliterated different from uncensored?

Uncensored is the general word. Abliterated names one method: editing the weights of an existing model. Dolphin and Hermes reach a similar result by training on data without refusals.

Does abliteration make a model worse?

Sometimes slightly: answers can be a little less careful than the original. Newer versions (often marked v2) are close to the original model in quality.

Which abliterated model should I try first?

On an 8 GB iPhone, Qwen3 4B Abliterated or Gemma 3 4B Abliterated. On a 12 GB iPhone or an iPad, Llama 3.1 8B Instruct Abliterated or Qwen3 8B Abliterated.