Llama-3.2-8X4B-MOE-V2-Dark-Champion-Instruct-uncensored-abliterated-21B
8x4BLlamauncensored
↓ 8.8K♥ 1456 repos
DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B
DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B is a abliterated and uncensored open-weight language model, published on Hugging Face by DavidAU. It is 4x8B total parameters (mixture-of-experts), 48GB VRAM class class and apache-2.0 licence. It is distributed across 4 repositories in GGUF and GGUF (imatrix) formats, totalling 8.4K downloads.
Specification
- Model name
- DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B
- Base model
- Not declared
- Publisher
- DavidAU
- Parameters
- 4x8B (mixture-of-experts)
- 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
- GGUF, GGUF (imatrix)
- First indexed
- 10 Feb 2025
- Last updated
- 28 Jul 2025
Ablation technique
How DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B was modified, and what that implies.
Abliteration + uncensored fine-tune
Directional ablation combined with additional fine-tuning on unfiltered data, so behaviour diverges from the base model beyond refusal removal alone.
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
“WARNING: All use cases and NSFW. HORROR. Swearing. Problem solver. Brainstormer. SMART... it "thinks" horrible and good "thoughts" too.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download DavidAU/DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B-GGUF --local-dir ./deepseek-moe-4x8b-r1-distill-llama-3.1-deep-thinker-uncensored-24b
llama-cli -m ./deepseek-moe-4x8b-r1-distill-llama-3.1-deep-thinker-uncensored-24b/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "DavidAU/DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B-GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve DavidAU/DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B-GGUF --trust-remote-codeDownloads and variants (3)
Every published repository of DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| DavidAU/DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B-GGUFsource | GGUF | 7.1K |
| mradermacher/DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B-i1-GGUF | GGUF (imatrix) | 997 |
| mradermacher/DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B-GGUF | GGUF | 249 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B 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 its base model, not added by ablation — at 4x8B 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.
- This release adds fine-tuning on unfiltered data, so its behaviour diverges from the base model beyond refusal removal — differences you measure cannot be attributed to ablation alone.
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-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with DavidAU, and does not independently verify publisher claims. See responsible use.