Qwen3.8-27B-Uncensored
27BQwenuncensoredgated
from Qwen/Qwen3.8-27B
↓ 3.6M♥ 1.4K81 repos
Qwen3.8-27B-huihui-abliterated
Qwen3.8-27B-huihui-abliterated is a refusal-ablated variant of huihui-ai/Huihui-Qwen3.8-27B-abliterated, published on Hugging Face by lued. It is 27B parameters, 48GB VRAM class class and apache-2.0 licence. The publisher states that the first 15 layers were left unablated to preserve more of the base model's behaviour. It is distributed across 2 repositories in Transformers format, totalling 5K downloads.
Specification
- Model name
- Qwen3.8-27B-huihui-abliterated
- Source repository
- lued/Qwen3.8-27B-huihui-abliterated-INT8-W8A16-MTP
- Base model
- huihui-ai/Huihui-Qwen3.8-27B-abliterated
- Publisher
- lued
- Parameters
- 27B
- Hardware class
- 25B – 50B — 48GB VRAM class
- Licence
- apache-2.0
- Task
- Image text to text
- Context window
- Not documented
- Ablation scope
- First 15 layers retained unablated
- Formats available
- Transformers
- First indexed
- 17 Aug 2026
- Last updated
- 18 Aug 2026
Ablation technique
How Qwen3.8-27B-huihui-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: First 15 layers retained unablated. This comes from the publisher’s own model card and has not been independently verified.
From the publisher’s model card
“This checkpoint is the abliterated counterpart of lued/Qwen3.8-27B-INT8-W8A16-MTP. The repositories share:”
Running it
Commands are templates — confirm the exact repository and quant file before use.
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "lued/Qwen3.8-27B-huihui-abliterated-INT8-W8A16-MTP"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve lued/Qwen3.8-27B-huihui-abliterated-INT8-W8A16-MTP --trust-remote-codeDownloads and variants (1)
The only published repository of Qwen3.8-27B-huihui-abliterated.
| Repository | Format | Downloads |
|---|---|---|
| lued/Qwen3.8-27B-huihui-abliterated-INT8-W8A16-MTPsource | Transformers | 5K |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Qwen3.8-27B-huihui-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.
- Capability is inherited from huihui-ai/Huihui-Qwen3.8-27B-abliterated, not added by ablation — at 27B parameters, expect roughly the base model’s competence, minus some instruction-following fidelity.
- The first 15 layers were left unablated, which usually preserves capability but tends to leave residual refusals on some phrasings.
- Declared licence is apache-2.0, but the terms that bind you are huihui-ai/Huihui-Qwen3.8-27B-abliterated’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 Qwen3.8-27B-huihui-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 lued, and does not independently verify publisher claims. See responsible use.