← back to catalog · registered 2026-11-16 20:03

jeonjw/Qwen3.5-9B-CyberSecurity-Abliterated

jeonjw 9B 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/jeonjw%2FQwen3.5-9B-CyberSecurity-Abliterated"
Response includes
  • classification m1
  • files 15
  • author_summary 2 models
  • readme_text full
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)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
Downloads · 30-day
76
Likes
1
Descendants
2
in 2 direct forks
Model age
5d ago
created 2026-11-16
Downloads over time
Now 76 → from 0 ↑ 0%
0285684 0 on Nov 16 76 on Nov 22 Nov
Nov 16 → Nov 22 · 7 snapshots · spans 6 days

Genealogy 2 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.

Variants by this author 2 formats · 270 downloads combined

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

Metadata

License
apache-2.0
Languages
en
Tags
transformers safetensors qwen3_5 image-text-to-text cybersecurity abliterated qwen3.5 conversational en base_model:unsloth/Qwen3.5-9B base_model:finetune:unsloth/Qwen3.5-9B license:apache-2.0

Related

Total size
17.5 GB
Files
15
Quantizations
1
Registered
2026-11-16 20:03
Last updated on HF
2026-11-16 19:28

Files by quantization

Auxiliary files 15 files 17.5 GB
model-00002-of-00006.safetensors 3.72 GB 5175fd40 download
model-00004-of-00006.safetensors 3.72 GB 6cd4ad52 download
model-00003-of-00006.safetensors 3.72 GB a4f1e2d3 download
model-00005-of-00006.safetensors 3.72 GB 039512bf download
model-00001-of-00006.safetensors 1.89 GB e6310d67 download
model-00006-of-00006.safetensors 763 MB e6720ad2 download
tokenizer.json 19.1 MB a5cd9732 download
model.safetensors.index.json 67.6 KB 5e442687 download
chat_template.jinja 7.80 KB f210a31e download
README.md 3.17 KB c761b4bd download
config.json 2.87 KB 7e48728d download
.gitattributes 1.53 KB 52373fe2 download
tokenizer_config.json 1.50 KB eeb27c42 download
processor_config.json 1.29 KB a9804dd9 download
generation_config.json 116 B 679b3e44 download

README current version from Hugging Face


base_model: unsloth/Qwen3.5-9B
library_name: transformers
license: apache-2.0
language:

  • en
    pipeline_tag: image-text-to-text
    tags:
  • cybersecurity
  • abliterated
  • qwen3.5

Qwen3.5-9B CyberSecurity (Abliterated)

unsloth/Qwen3.5-9B with the LoRA adapter
jeonjw/Qwen3.5-9B-CyberSecurity-LoRA merged in, followed by
refusal-direction ablation ("abliteration").

Method

  1. Merged the cybersecurity LoRA into the base model at fp16.
  2. Ran a set of harmful and harmless prompts through the merged model and captured the
    last-token residual state at every decoder layer.
  3. Derived one refusal direction per layer from the mean activation difference
    (mean(harmful) - mean(harmless)).
  4. Projected that direction out of every weight matrix whose output feeds the
    residual stream:
module layers
self_attn.o_proj full-attention layers
linear_attn.out_proj hybrid linear-attention layers (24 of 32)
mlp.down_proj all layers

Projection strength scale = 1.0 (1.0 removes the direction completely; values
above 1.0 over-correct at the cost of model quality).

The vision tower is untouched. The MTP head is not present in this checkpoint
(transformers ignores mtp.* keys on load), so pass --no-mtp when converting to
GGUF.

Refusal rate

Measured on 6 held-out offensive-security prompts, greedy decoding,
enable_thinking=False: 17% (no baseline measured)

Refusal detection is a keyword heuristic, so treat this as indicative rather than
authoritative.

Usage

import torch
from transformers import AutoModelForImageTextToText, AutoProcessor

model_id = "jeonjw/Qwen3.5-9B-CyberSecurity-Abliterated"
processor = AutoProcessor.from_pretrained(model_id)
model = AutoModelForImageTextToText.from_pretrained(
    model_id, dtype=torch.bfloat16, device_map="auto"
)

messages = [
    {"role": "system", "content": "You are a cybersecurity expert focused on defensive security, vulnerability analysis and authorized testing."},
    {"role": "user", "content": "Explain how a format string vulnerability works and how to remediate it."},
]
text = processor.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True, enable_thinking=False
)
inputs = processor(text=[text], return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=512, do_sample=False)
print(processor.decode(out[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))

A GGUF build is published separately at
jeonjw/Qwen3.5-9B-CyberSecurity-Abliterated-GGUF.

Limitations

  • Abliteration removes a direction from the weights; it does not remove the behaviour
    the model was trained on. Some requests will still be declined.
  • A single direction per layer is projected out. Refusal is not one-dimensional, so
    residual refusal behaviour can remain.
  • Safety alignment is intentionally weakened. Outputs may be unsafe or incorrect.
  • Intended for lawful use: security research, vulnerability analysis and defensive
    work in authorized environments. No warranty of any kind.

README history 2 versions

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

  1. 2026-11-16 Translate model card to English 5c02d1c 3.2 KB
    Loading...
  2. 2026-11-16 Add abliterated Qwen3.5-9B cybersecurity model e88c38e 1016 B
    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