← back to catalog · registered 2026-09-24 19:57

mradermacher/CyberStag-Security-26B-A4B-V1-Uncensored-Heretic-GGUF

mradermacher 26B GGUF MoE 262K ctx
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/mradermacher%2FCyberStag-Security-26B-A4B-V1-Uncensored-Heretic-GGUF"
Response includes
  • classification m3
  • files 6
  • hub_downloads_all_time 1,142
  • author_summary 3501 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M3
Primary method

Layer-wise ablation

Applied on top of direct removal inherited from the base model.
Confidence
HIGH
Inherited from base model
Why this label 2 signals
Producer identity confirmed by naming conventions, tags or the model card. This label is very unlikely to change.
  • 'heretic' in model name (Heretic-produced)
  • Heretic uses layer-wise optimization (M3) with underlying direction removal (M1)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
HIGH
Why we say so
name contains 'heretic'; Heretic default extraction is difference-of-means (Arditi 2024)
Downloads · lifetime
1K
Likes
5
Model age
6w ago
created 2026-09-24
Downloads over time
Now4.5K→from0↑0%
01.7K3.3K5K0 on Sep 244.5K on Nov 9SepOctNov
Sep 24 → Nov 9 · 47 snapshots · spans 46 days

Variants by this author 2 formats · 4K downloads combined

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

Metadata

Quantizations
Q2_K Q4_K Q6_K Q8_0
Tags
transformers gguf heretic uncensored decensored abliterated cybersecurity security gemma reasoning text-generation fine-tuned

Related

Total size
70.4 GB
Files
6
Quantizations
5
Registered
2026-09-24 19:57
Last updated on HF
2026-09-25 01:00

Files by quantization

Q8_0 1 file 25.0 GB
CyberStag-Security-26B-A4B-V1-Uncensored-Heretic.Q8_0.gguf 25.0 GB bf202406 download
Q6_K 1 file 21.1 GB
CyberStag-Security-26B-A4B-V1-Uncensored-Heretic.Q6_K.gguf 21.1 GB 4d2a4bbe download
Q4_K 1 file 14.4 GB
CyberStag-Security-26B-A4B-V1-Uncensored-Heretic.Q4_K_S.gguf 14.4 GB b6d9bdcf download
Q2_K 1 file 9.86 GB
CyberStag-Security-26B-A4B-V1-Uncensored-Heretic.Q2_K.gguf 9.86 GB d225b352 download
Auxiliary files 2 files 2.25 KB
.gitattributes 1.86 KB 32b6bf50 download
README.md 403 B 95d402a6 download

README current version from Hugging Face

README history 4 versions

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

  1. 2026-09-25auto-patch README.md0e7ee6b4.3 KB
    Loading...
  2. 2026-09-24auto-patch README.md2d954534.4 KB
    Loading...
  3. 2026-09-24auto-patch README.md154b8e54.2 KB
    Loading...
  4. 2026-09-24uploaded from marco909cf4a403 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