DeepSeek-V4.1-Flash-Uncensored
DeepSeekuncensored
from deepseek-ai/DeepSeek-V4.1-Flash
↓ 68.9K♥ 82752 repos
DeepSeek-V4.1-Flash-Abliterated
DeepSeek-V4.1-Flash-Abliterated is a refusal-ablated variant of deepseek-ai/DeepSeek-V4.1-Flash, published on Hugging Face by s-zaizen. It is mit licence. It is distributed across 12 repositories in Transformers, GGUF, EXL3 and FP8 formats, totalling 90.7K downloads.
Specification
- Model name
- DeepSeek-V4.1-Flash-Abliterated
- Source repository
- s-zaizen/DeepSeek-V4.1-Flash-Abliterated
- Base model
- deepseek-ai/DeepSeek-V4.1-Flash
- Publisher
- s-zaizen
- Parameters
- Not determinable from name
- VRAM required
- Unknown — no published parameter count
- Hardware class
- Unknown
- Licence
- mit
- Task
- Image text to text
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, EXL3, FP8
- First indexed
- 10 Sept 2026
- Last updated
- 27 Oct 2026
Ablation technique
How DeepSeek-V4.1-Flash-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 a community abliterated conversion of DeepSeek-V4.1-Flash, published by s-zaizen. It is not an official DeepSeek release.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download audreyt/DeepSeek-V4.1-Flash-Abliterated-GGUF --local-dir ./deepseek-v4.1-flash-abliterated
llama-cli -m ./deepseek-v4.1-flash-abliterated/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "s-zaizen/DeepSeek-V4.1-Flash-Abliterated"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve s-zaizen/DeepSeek-V4.1-Flash-Abliterated --trust-remote-codeDownloads and variants (11)
Every published repository of DeepSeek-V4.1-Flash-Abliterated, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| audreyt/DeepSeek-V4.1-Flash-Abliterated-GGUF | GGUF | 88.4K |
| s-zaizen/DeepSeek-V4.1-Flash-Abliteratedsource | Transformers | 816 |
| bigfat/DeepSeek-V4.1-Flash-Abliterated-GGUF | GGUF | 421 |
| securepeak/DeepSeek-V4.1-Flash-Abliterated | Transformers | 399 |
| distributedcognition/DeepSeek-V4.1-Flash-abliterated | Transformers | 229 |
| rax4901/DeepSeek-V4.1-Flash-Abliterated | Transformers | 218 |
| patto91/DeepSeek-V4.1-Flash-Abliterated | Transformers | 94 |
| mIn-MAXsystem-Htz/DeepSeek-V4.1-Flash-Abliterated | Transformers | 75 |
| UltimateIntent/DeepSeek-V4.1-Flash-Abliterated-EXL3-3.51bpw | EXL3 | 0 |
| dealignai/DeepSeek-V4.1-Flash-ABLITERATED-FP8 | FP8 | 0 |
| soumabhanoct/DeepSeek-V4.1-Flash-Abliterated-Q2-H200 | Transformers | 0 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- DeepSeek-V4.1-Flash-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 mit, but the terms that bind you are deepseek-ai/DeepSeek-V4.1-Flash’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 DeepSeek-V4.1-Flash-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 s-zaizen, and does not independently verify publisher claims. See responsible use.