CyberNeurova-DeepSeek-V4-Flash-abliterated
DeepSeek
from deepseek-ai/DeepSeek-V4-Flash
↓ 10.9K♥ 615 repos
DeepSeek-V4-Flash-abliterated-ds4
DeepSeek-V4-Flash-abliterated-ds4 is a refusal-ablated variant of deepseek-ai/DeepSeek-V4-Flash, published on Hugging Face by huihui-ai. It is mit licence. It is distributed across 7 repositories in GGUF format, totalling 229K downloads.
Specification
- Model name
- DeepSeek-V4-Flash-abliterated-ds4
- Source repository
- huihui-ai/Huihui-DeepSeek-V4-Flash-abliterated-ds4-GGUF
- Base model
- deepseek-ai/DeepSeek-V4-Flash
- Publisher
- huihui-ai
- Parameters
- Not determinable from name
- Hardware class
- Unknown
- Licence
- mit
- Task
- Unspecified
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- GGUF
- First indexed
- 26 May 2026
- Last updated
- 13 Aug 2026
Ablation technique
How DeepSeek-V4-Flash-abliterated-ds4 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 an uncensored version of deepseek-ai/DeepSeek-V4-Flash created with abliteration.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download huihui-ai/Huihui-DeepSeek-V4-Flash-abliterated-ds4-GGUF --local-dir ./deepseek-v4-flash-abliterated-ds4
llama-cli -m ./deepseek-v4-flash-abliterated-ds4/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "huihui-ai/Huihui-DeepSeek-V4-Flash-abliterated-ds4-GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve huihui-ai/Huihui-DeepSeek-V4-Flash-abliterated-ds4-GGUF --trust-remote-codeDownloads and variants (6)
Every published repository of DeepSeek-V4-Flash-abliterated-ds4, including quantised re-releases by other authors.
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- DeepSeek-V4-Flash-abliterated-ds4 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-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-Flash-abliterated-ds4 is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with huihui-ai, and does not independently verify publisher claims. See responsible use.