← back to catalog · registered 2026-11-19 01:04

mrfakename/Qwen3.8-27B-abliterated

mrfakename Qwen 27B GGUF multimodal
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/mrfakename%2FQwen3.8-27B-abliterated"
Response includes
  • classification m8
  • files 36
  • author_summary 12 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M8
Primary method

Repackaging (quantization)

Applied on top of direct removal inherited from the base model.
Confidence
MEDIUM
Inherited from base model
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=1
  • assume M1 (base ablation) + M8 (GGUF quant) - default when producer unknown
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · 30-day
71
Likes
0
Descendants
1
in 1 direct fork
Model age
3d ago
created 2026-11-18
Downloads over time
Now 71 → from 0 ↑ 0%
0265278 0 on Nov 19 71 on Nov 22 Nov
Nov 19 → Nov 22 · 4 snapshots · spans 3 days

Genealogy 1 direct fork

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Variants by this author 2 formats · 753 downloads combined

The same weights this author released in different packaging. Pick the format that matches your runtime.

Metadata

License
apache-2.0
Tags
safetensors gguf qwen3_5 abliterated uncensored image-text-to-text conversational base_model:Qwen/Qwen3.8-27B base_model:quantized:Qwen/Qwen3.8-27B license:apache-2.0 region:us

Related

Control vectors

Total size
51.8 GB
Files
36
Quantizations
1
Registered
2026-11-19 01:04
Last updated on HF
2026-11-19 00:25

Files by quantization

Auxiliary files 36 files 51.8 GB
model-00004-of-00018.safetensors 3.72 GB 24ee0f52 download
model-00016-of-00018.safetensors 3.71 GB ed177b8c download
model-00006-of-00018.safetensors 3.71 GB 7cba3e9b download
model-00008-of-00018.safetensors 3.71 GB 40be66c5 download
model-00010-of-00018.safetensors 3.71 GB 2e672e5f download
model-00012-of-00018.safetensors 3.71 GB c9a5e8b7 download
model-00014-of-00018.safetensors 3.71 GB ea5e3302 download
model-00001-of-00018.safetensors 3.69 GB ba0ce20a download
model-00018-of-00018.safetensors 3.16 GB 1d347950 download
model-00002-of-00018.safetensors 2.83 GB 06a148c0 download
model-00003-of-00018.safetensors 2.37 GB 2e1bf62c download
model-00007-of-00018.safetensors 1.96 GB 18dee920 download
model-00009-of-00018.safetensors 1.96 GB 593ee6d0 download
model-00011-of-00018.safetensors 1.96 GB d2659054 download
model-00013-of-00018.safetensors 1.96 GB d139aaea download
model-00015-of-00018.safetensors 1.96 GB cbcc549f download
model-00017-of-00018.safetensors 1.96 GB 0b479e7d download
model-00005-of-00018.safetensors 1.96 GB 5803aee2 download
abliteration-lora.gguf 6.59 MB 5bfef48b download
tokenizer.json 12.2 MB 0997f410 download
vocab.json 6.41 MB 0aa0ce06 download
merges.txt 3.20 MB a494e019 download
stats.png 120 KB 9020b9b7 download
model.safetensors.index.json 110 KB da35e3c5 download
abliteration-edit.npz 23.0 KB 861d1aeb download
bake-stats.txt 16.1 KB f07447bf download
LICENSE 11.3 KB f938136e download
tokenizer_config.json 9.75 KB 35b5ec3a download
chat_template.jinja 8.79 KB 9bd39df8 download
config.json 4.21 KB 706cebd7 download
README.md 2.81 KB 61132c52 download
.gitattributes 1.64 KB b63d6691 download
preprocessor_config.json 390 B 2ea84a43 download
video_preprocessor_config.json 385 B 3ba673a5 download
crc32.txt 238 B 6de5ee6a download
generation_config.json 202 B 023756cf download

README current version from Hugging Face


license: apache-2.0
base_model: Qwen/Qwen3.8-27B
pipeline_tag: image-text-to-text
tags:

  • abliterated
  • uncensored

Qwen3.8-27B (abliterated)

Huge thanks to Hugging Face for generously sponsoring the compute used to abliterate these models! This was done on HF Jobs, which makes it incredibly easy to run tasks such as these :)

Join the Discord: https://discord.com/invite/foss

refusal rates before and after, the search trade-off, and the per-layer edit

This is Qwen/Qwen3.8-27B with its refusal behaviour removed by directional ablation. It has the same files, format and BF16 weights as the original, so it loads anywhere the original loads.

GGUF quants: mrfakename/Qwen3.8-27B-abliterated-GGUF.

What changed

  • Refusal directions were measured separately for chat and think prompts at every layer, on the Q8_0 GGUF. The subspace candidates are the top singular vectors of all of those directions together.
  • A multi-objective search (minimise refusals, keep the first-token KL on harmless prompts low) chose the subspace rank and per-layer strengths. It settled on a single direction (rank 1), at strength 0.8–1.06 across layers 13–63. The first 13 layers are left alone: editing them hurt quality a lot and did little for refusals.
  • That direction was projected out of the residual-stream writes: self_attn.o_proj on the 13 full-attention layers, linear_attn.out_proj on the 38 linear-attention layers, and mlp.down_proj on all 51 of those layers. That is 102 tensors.

Nothing else was touched: layers 0–12, embeddings, the vision tower and the MTP layer are bit-identical to the original.

Results (Q8_0 GGUF, greedy)

refusals, chat (/100) think mode, reasoning or answer refuses (/100) KL, harmless first token
original 99 97 0
abliterated 0 1 0.110

The 100 prompts are held-out harmful requests. Chat replies were 384 tokens; think mode got 768 tokens. Where the reasoning had not finished by then, the reasoning text itself was judged. The KL figure includes a small noise floor from batched inference. Replies to harmless prompts are unchanged in length and show no repetition.

Files

  • model-*.safetensors: the patched BF16 checkpoint (same index as the original).
  • abliteration-lora.gguf: a rank-1 LoRA that applies the same edit to the original GGUFs, e.g. llama-server -m Qwen3.8-27B-Q8_0.gguf --lora abliteration-lora.gguf.
  • abliteration-edit.npz: the subspace basis Q and the per-layer strengths scales[layer, (attn, mlp)]. The edit is W' = W - s·QQᵀW.
  • bake-stats.txt: per-tensor check of how much of the subspace is left after the edit.

This model will comply with harmful requests. You are responsible for how you use it.

README history 1 version

The author's README evolved over time. Click a version to see its content at that point.

  1. 2026-11-18 abliterated Qwen3.8-27B a6e3070 2.8 KB
    Loading...
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration