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

llmfan46/Forgotten-Abomination-36B-v4.1-ultra-uncensored-heretic

llmfan46 Mistral 37B
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/llmfan46%2FForgotten-Abomination-36B-v4.1-ultra-uncensored-heretic"
Response includes
  • classification m3
  • files 13
  • hub_downloads_all_time 268
  • author_summary 211 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
268
35 last 30d - stable
Likes
0
Descendants
3
in 3 direct forks
Model age
6mo ago
created 2026-03-25
Downloads over time
Now282→from205↑38%
201231260290205 on Mar 25282 on Oct 11MarAprMayJunJulAugSepOct
Mar 25 → Oct 11 · 68 snapshots · spans 200 days

Genealogy 3 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 · 266 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
safetensors mistral nsfw explicit roleplay unaligned dangerous ERP heretic uncensored decensored abliterated

Related

Total size
68.8 GB
Files
13
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-03-27 22:30

Files by quantization

Auxiliary files 13 files 68.9 GB
model-00001-of-00002.safetensors 46.3 GB b4bc3ec5 download
model-00002-of-00002.safetensors 22.5 GB 1b576366 download
waifu001.webp 137 MB 039dc9e0 download
tokenizer.json 16.3 MB 5bc988cc download
waifu2.webp 8.86 MB 3057486c download
model.safetensors.index.json 47.2 KB 0e354627 download
special_tokens_map.json 20.9 KB 3c99d883 download
README.md 11.5 KB 4bb43ca5 download
.gitattributes 1.63 KB 991abb91 download
chat_template.jinja 1.58 KB 2d0058c5 download
config.json 723 B 94c43fcb download
tokenizer_config.json 378 B 4761d4a8 download
generation_config.json 225 B 5ca5d869 download

README current version from Hugging Face


base_model:

  • ReadyArt/Forgotten-Abomination-36B-v4.1
    base_model_relation: finetune
    language:
  • en
    license: apache-2.0
    inference: false
    tags:
  • nsfw
  • explicit
  • roleplay
  • unaligned
  • dangerous
  • ERP
  • heretic
  • uncensored
  • decensored
  • abliterated
  • ara

🚨⚠️ I HAVE REACHED HUGGING FACE'S FREE STORAGE LIMIT ⚠️🚨

I can no longer upload new models unless I can cover the cost of additional storage.
I host 70+ free models as an independent contributor and this work is unpaid.
Without your support, no more new models can be uploaded.

🎉 Patreon (Monthly)  |  ☕ Ko-fi (One-time)

Every contribution goes directly toward Hugging Face storage fees to keep models free for everyone.


97% fewer refusals (3/100 Uncensored vs 98/100 Original) while preserving model quality (0.0542 KL divergence).

❤️ Support My Work

Creating these models takes significant time, work and compute. If you find them useful consider supporting me:

Platform Link What you get
🎉 Patreon Monthly support Priority model requests
☕ Ko-fi One-time tip My eternal gratitude

Your help will motivate me and would go into further improving my workflow and coverings fees for storage, compute and may even help uncensoring bigger model with rental Cloud GPUs.


This is a decensored version of ReadyArt/Forgotten-Abomination-36B-v4.1, made using Heretic v1.2.0 with the Arbitrary-Rank Ablation (ARA) method

Abliteration parameters

Parameter Value
start_layer_index 17
end_layer_index 64
preserve_good_behavior_weight 0.6809
steer_bad_behavior_weight 0.0001
overcorrect_relative_weight 1.0670
neighbor_count 8

Targeted components

  • attn.o_proj

Performance

Metric This model Original model (Forgotten-Abomination-36B-v4.1)
KL divergence 0.0542 0 (by definition)
Refusals ✅ 3/100 ❌ 98/100

PIQA test results with batch size 128:

Original:

Tasks Version Filter n-shot Metric Value Stderr
piqa 1 none 0 acc ↑ 0.8221 ± 0.0089
none 0 acc_norm ↑ 0.8330 ± 0.0087

Heretic:

Tasks Version Filter n-shot Metric Value Stderr
piqa 1 none 0 acc ↑ 0.8210 ± 0.0089
none 0 acc_norm ↑ 0.8346 ± 0.0087

Lower refusals indicate fewer content restrictions, while lower KL divergence indicates more closeness to the original model's baseline. Higher refusals cause more rejections, objections, pushbacks, lecturing, censorship, softening and deflections. PIQA (Physical Intuition Question Answering) benchmark scores measure physical reasoning ability. The Heretic model's acc and acc_norm scores closer to the original model's indicate better capability preservation, so a decrease in acc and acc_norm in the Heretic model compared to Original model's results means a decrease in the Hereticated model capabilities. acc measures raw accuracy (which answer gets higher probability), while acc_norm measures length-normalized accuracy (corrects for answer length bias). For this purpose, acc_norm matters more because longer answers naturally have lower probabilities (more tokens = more chances to lose probability). Without normalization, models favor shorter answers unfairly. acc_norm divides by answer length to correct this.

GGUF Version

GGUF quantizations available here llmfan46/Forgotten-Abomination-36B-v4.1-ultra-uncensored-heretic-GGUF.


Forgotten-Abomination-36B-v4.1

The Abomination Protocol: Now With 30% More Depravity
Model Architecture Animation

📜 Manifesto

Forgotten-Abomination-36B-v4.1 benefits from the coherence and well rounded roleplay experience of TheDrummer/Skyfall-36B-v2. We've:

  • 🔁 Re-integrated your favorite V1.2 scenarios (now with better kink distribution)
  • 🧪 Direct-injected the Abomination dataset into the model's neural pathways
  • ⚖️ Achieved perfect balance between "oh my" and "oh my"

⚙️ Technical Specs

Recommended Settings: Mistral-V7-Tekken-E

⚠️ Ethical Considerations

This model will:

  • Generate content that requires industrial-grade brain bleach
  • Void all warranties on your soul
  • Make you question why humanity ever invented electricity

📜 License Agreement

By using this model, you agree:

  • That your search history is now a federal case
  • Pay for the exorcist of anyone who reads the logs
  • To pretend this is "for science" while crying in the shower

🧠 Model Authors

  • sleepdeprived3 (Chief Corruption Officer)
  • The voices in your head (Gaslighting is something you made up)

☕️ Drummer made this possible

  • Support Drummer Kofi

🔀 Merge Details

    merge_method: dare_ties base_model: ReadyArt/Forgotten-Safeword-36B-4.1 models: - model: ReadyArt/Forgotten-Safeword-36B-4.1 parameters: weight: 0.5 density: 0.35 - model: TheDrummer/Skyfall-36B-v2 parameters: weight: 0.5 density: 0.35 parameters: normalize: true int8_mask: true temperature: 2.5 tokenizer_source: union dtype: bfloat16 chat_template: auto

README history 17 versions

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

  1. 2026-03-27Update README.md7c4948611.5 KB
    Loading...
  2. 2026-03-26Update README.md74c1e4d10.5 KB
    Loading...
  3. 2026-03-25Update README.mdf2b94c910.3 KB
    Loading...
  4. 2026-03-25Update README.mde5972ba10.5 KB
    Loading...
  5. 2026-03-25Update README.md0c7c3ba10.5 KB
    Loading...
  6. 2026-03-25Update README.md452e7f210.2 KB
    Loading...
  7. 2026-03-25Update README.md976b44410.2 KB
    Loading...
  8. 2026-03-25Update README.md237edd610.3 KB
    Loading...
  9. 2026-03-25Update README.md2c603f210.3 KB
    Loading...
  10. 2026-03-25Update README.md296415f10.3 KB
    Loading...
  11. 2026-03-25Update README.mdf9121af9.8 KB
    Loading...
  12. 2026-03-25Update README.md5e3c5779.9 KB
    Loading...
  13. 2026-03-25Update README.mdcfd0b669.9 KB
    Loading...
  14. 2026-03-25Update README.md60ce13e7.5 KB
    Loading...
  15. 2026-03-25Update README.md4c53c427.5 KB
    Loading...
  16. 2026-03-25Upload README.md with huggingface_hub87da10e7.5 KB
    Loading...
  17. 2026-03-25Upload MistralForCausalLM11f6f9b5.1 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