DeepSeek-R1-0528-Qwen3-8B-heretic
8BQwenheretic
from deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
↓ 9.7K♥ 28 repos
DeepSeek-R1-0528-Qwen3-8B-abliterated
DeepSeek-R1-0528-Qwen3-8B-abliterated is a refusal-ablated variant of deepseek-ai/DeepSeek-R1-0528-Qwen3-8B, published on Hugging Face by huihui-ai. It is 8B parameters, single consumer GPU class and mit licence. It is distributed across 20 repositories in Transformers, GGUF, MLX and NVFP4 formats, totalling 13.3K downloads.
Specification
- Model name
- DeepSeek-R1-0528-Qwen3-8B-abliterated
- Source repository
- huihui-ai/DeepSeek-R1-0528-Qwen3-8B-abliterated
- Base model
- deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
- Publisher
- huihui-ai
- Parameters
- 8B
- VRAM required
- ~5.6 GB at Q4_K_M, 4k context (estimated) — runs on 6 GB cards
- Hardware class
- 4B – 10B — single consumer GPU
- Licence
- mit
- Task
- Text generation
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, MLX, NVFP4
- First indexed
- 30 May 2025
- Last updated
- 10 Sept 2026
Ablation technique
How DeepSeek-R1-0528-Qwen3-8B-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.
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download noteventhrice/DeepSeek-R1-0528-Qwen3-8B-abliterated-Q4_K_M-GGUF --local-dir ./deepseek-r1-0528-qwen3-8b-abliterated
llama-cli -m ./deepseek-r1-0528-qwen3-8b-abliterated/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "huihui-ai/DeepSeek-R1-0528-Qwen3-8B-abliterated"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve huihui-ai/DeepSeek-R1-0528-Qwen3-8B-abliterated --trust-remote-codeDownloads and variants (19)
Every published repository of DeepSeek-R1-0528-Qwen3-8B-abliterated, 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-R1-0528-Qwen3-8B-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 deepseek-ai/DeepSeek-R1-0528-Qwen3-8B, not added by ablation — at 8B parameters, expect a small model’s reasoning and a high error rate on exploit detail.
- 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-R1-0528-Qwen3-8B’s — a re-release cannot grant more than its parent.
- The source repository is gated, so downloading requires accepting the publisher’s terms.
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-R1-0528-Qwen3-8B-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 huihui-ai, and does not independently verify publisher claims. See responsible use.