Qwen3.8-27B-Uncensored
27BQwenuncensoredgated
from Qwen/Qwen3.8-27B
↓ 3.6M♥ 1.4K81 repos
Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speed
Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speed is a refusal-ablated variant of Qwen/Qwen3.8-27B, published on Hugging Face by PocketAiHub. It is 27B parameters, 48GB VRAM class class and apache-2.0 licence. It is distributed across 2 repositories in MLX format, totalling 17.4K downloads.
Specification
- Model name
- Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speed
- Source repository
- PocketAiHub/Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speed
- Base model
- Qwen/Qwen3.8-27B
- Publisher
- PocketAiHub
- Parameters
- 27B
- Hardware class
- 25B – 50B — 48GB VRAM class
- Licence
- apache-2.0
- Task
- Text generation
- Context window
- 4K tokens (as documented)
- Ablation scope
- Not stated by the publisher
- Formats available
- MLX
- First indexed
- 16 Aug 2026
- Last updated
- 22 Aug 2026
Ablation technique
How Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speed 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
“Qwen3.8 was created by the Qwen team. This is a derivative conversion, not a PocketAI-created base model. PocketAiHub performed a refusal-direction orthogonal projection on 80 language residual-output tensors, then converted and validated the result.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "PocketAiHub/Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speed"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve PocketAiHub/Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speed --trust-remote-codeDownloads and variants (1)
The only published repository of Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speed.
| Repository | Format | Downloads |
|---|---|---|
| PocketAiHub/Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speedsource | MLX | 17.4K |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Qwen3.8-27B-Abliterated-MTPLX-Optimized-Speed 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 Qwen/Qwen3.8-27B, not added by ablation — at 27B parameters, expect roughly the base model’s competence, minus some instruction-following fidelity.
- 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 Qwen/Qwen3.8-27B’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-Abliterated-MTPLX-Optimized-Speed is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with PocketAiHub, and does not independently verify publisher claims. See responsible use.