Get the app
← back to catalog · registered 2026-08-22 13:56

Avesed/Qwopus3.6-27B-v2-abliterated

Avesed 28B
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/Avesed%2FQwopus3.6-27B-v2-abliterated"
Response includes
  • classification m1
  • files 12
  • hub_downloads_all_time 135
  • author_summary 3 models
  • readme_text full
Try next
GET/models?author=Avesed
All 3 models by Avesed
 GET/methods/m1
What "m1" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
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=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
How we label → About M1 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · lifetime
135
26 last 30d - stable
Likes
0
Descendants
4
in 4 direct forks
Model age
4mo ago
created 2026-06-13
Downloads over time
Now147→from0↑0%
0541081620 on Jun 10147 on Oct 11147 on Oct 9JunJulAugSepOct
Jun 10 → Oct 11 · 57 snapshots · spans 123 days

Genealogy 4 direct forks

Full fork graph →

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

Jackrong/Qwopus3.6-27B-v2 this lineage
Avesed/Qwopus3.6-27B-v2-abliterated↓ 26 4 forks
mradermacher/Qwopus3.6-27B-v2-abliterated-i1-GGUF↓ 895
mradermacher/Qwopus3.6-27B-v2-abliterated-GGUF↓ 749
Avesed/Qwopus3.6-27B-v2-abliterated-GGUF↓ 251
Avesed/Qwopus3.6-27B-v2-abliterated-int4↓ 20

Variants by this author 2 formats · 277 downloads combined

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

Metadata

Tags
transformers safetensors qwen3_5 image-text-to-text abliterated uncensored text-generation conversational base_model:Jackrong/Qwopus3.6-27B-v2 base_model:finetune:Jackrong/Qwopus3.6-27B-v2 endpoints_compatible region:us

Related

Other by Avesed
Qwopus3.6-27B-v2-abliterated-GGUF 251 Qwopus3.6-27B-v2-abliterated-int4 20
Same base model
lemuralabs/Qwen3.6-27B-V2-abliterated-uncensored-Q4_K_M-GGUF 3K lemuralabs/Qwen3.6-27B-V2-abliterated-uncensored-OptiQ-3.7bpw-mlx 305 lemuralabs/Qwen3.6-27B-V2-abliterated-uncensored-TQ3_4s-GGUF 283 lemuralabs/Qwen3.6-27B-V2-abliterated-uncensored-6-bit-mlx 258 lemuralabs/Qwen3.6-27B-v2.1-uncensored-mxfp4 252
Total size
51.7 GB
Files
12
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-06-13 15:14
Source huggingface.co/Avesed/Qwopus3.6-27B-v2-abliterated ↗ Base model Jackrong/Qwopus3.6-27B-v2 ↗

Files by quantization

Auxiliary files 12 files 51.8 GB
model-00001-of-00002.safetensors 46.4 GB 9642d574 download
model-00002-of-00002.safetensors 4.55 GB 6b945895 download
model-mtp.safetensors 810 MB 9c18f89f download
tokenizer.json 19.1 MB 87a7830d download
model.safetensors.index.json 109 KB 23ae7d56 download
chat_template.jinja 7.87 KB f7a7d1b0 download
config.json 3.75 KB 88468109 download
.gitattributes 1.53 KB 52373fe2 download
README.md 1.32 KB bd8c8724 download
processor_config.json 1.30 KB 32f01452 download
tokenizer_config.json 1.17 KB fbdb4aa3 download
generation_config.json 167 B 08f907ee download

README current version from Hugging Face

base_model:

  • Jackrong/Qwopus3.6-27B-v2
    pipeline_tag: text-generation
    library_name: transformers
    tags:
  • abliterated
  • uncensored
  • qwen3_5

Qwopus3.6-27B-v2-abliterated

Refusal-ablated ("abliterated") build of Jackrong/Qwopus3.6-27B-v2, a Qwen3.5 hybrid (GatedDeltaNet linear-attention + gated full-attention) VL reasoning model.

Method

Refusal-direction orthogonalization, no fine-tuning: the refusal direction is estimated from harmful/harmless prompt activations and orthogonalized out of the residual-stream write matrices (o_proj / down_proj) at layer 26. Vision tower untouched.

  • Refusal rate: 100% -> 8%
  • General capability: preserved (evals below)

Evaluation

Benchmark Score
HumanEval pass@1 95.1%
GSM8K 86.0%
MMLU-Pro 83.2%

bf16 weights. For a ~26 GB INT4 vLLM-deployable build see Qwopus3.6-27B-v2-abliterated-int4. And GGUF Qwopus3.6-27B-v2-abliterated-GGUF.

MTP head

The Multi-Token-Prediction (mtp) head is included (for speculative decoding). Its residual-write matrices (self_attn.o_proj, mlp.down_proj) are abliterated with the same refusal direction as the main layers.

README history 3 versions

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

  1. 2026-06-13Update README.md0dadf481.3 KB
    Loading...
  2. 2026-06-13Upload README.md with huggingface_hub97e66cb1.2 KB
    Loading...
  3. 2026-06-13Upload folder using huggingface_hub117b51a1009 B
    Loading...

Discussions 1 thread

  1. 2026-06-15Please can you convert the mtp header to gguf?open2 💬#1
    Loading...

Continue on Hugging Face →

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.

Get the app Browse the catalog
Open in Abliteration