base_model:

• llmfan46/Omega-Evolution-27B-v2.0-ultra-uncensored-heretic
  base_model_relation: finetune
  tags:
• nsfw
• explicit
• roleplay
• unaligned
• dangerous
• ERP
• Other License
• heretic
• uncensored
• decensored
• abliterated
• ara
• nvfp4
  license: apache-2.0
  datasets:
• neuralmagic/calibration

Deaquay/Omega-Evolution-27B-v2.0-ultra-uncensored-heretic-NVFP4

This model Deaquay/Omega-Evolution-27B-v2.0-ultra-uncensored-heretic-NVFP4 was converted to NVFP4 format from llmfan46/Omega-Evolution-27B-v2.0-ultra-uncensored-heretic using llm-compressor.

(Includes a fixed chat template for qwen3.5+3.6 which allows for system messages anywhere, and replaces empty user messages with " " to make sending empty messages possible, fixing some problems that exists with agents and as an example extensions in sillytavern)

Original Model Card

97% fewer refusals (3/100 Uncensored vs 91/100 Original) while preserving model quality (0.1005 KL divergence).

❤️ Support My Work

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

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 zerofata/MS3.2-PaintedFantasy-Visage-v3-34B, made using Heretic v1.2.0 with the Arbitrary-Rank Ablation (ARA) method

Abliteration parameters

Targeted components

• attn.o_proj

Performance

PIQA test results with batch size 128:

Original:

Heretic:

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) a ~1,800 questions tests common-sense understanding of how the physical world works with benchmark scores to measure physical reasoning ability. The Heretic model's acc and acc_norm scores closer to the original model's indicate better capability preservation, a big decrease in acc and acc_norm in the Heretic model compared to Original model's results means a big 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.

MMLU test results with batch size 64:

Original:

Heretic:

MMLU - Massive Multitask Language Understanding, ~14,000 multiple-choice questions across 57 subjects (math, history, law, medicine, etc.).

GGUF Version

GGUF quantizations available here llmfan46/Omega-Evolution-27B-v2.0-ultra-uncensored-heretic-GGUF.

😈 OMEGA EVOLUTION V2.0 😈

⚠️ 27B Parameters ⚠️

⚠️ Thinking works (RP prompts) ⚠️

🔴 CLASSIFIED WARNINGS

• This is a hybrid construct of Safeword Omega Directive, Safeword Omega Darker, and Brisk Evolution v0.3.
• CONTENT WARNING: NSFW, Explicit, ERP, and Unaligned behavior are enabled by default.
• Dataset Revamp Took a sledgehammer to the dataset. Most formatting issues should be gone now.

⚙️ SYSTEM PARAMETERS

top_p 0.95

temp 0.9

🧪 ARCHITECTS

• GECFDO (Dataset Generation & Quants)
• Darkhn (Dataset Cleanup Tool)
• Sleep Deprived (Safeword Creator)
• FrenzyBiscuit (Brisk Evolution Creator)

🔥 LICENSE: APACHE 2.0 (WITH MORAL DISCLAIMER) 🔥
You accept full responsibility for corruption. You are 18+. The architects are not liable for the depravity you unleash.