base_model, library_name, language, license, tags, inference, pipeline_tag

base_model	library_name	language	license	tags	inference	pipeline_tag
hereticness/Heretic-Dirty-Alice-RP-NSFW-llama-3.2-1B UmbrellaInc/T-Virus_Delta.Strain-3.2-1B hereticness/heretic_FuseChat-Llama-3.2-1B-Instruct marcuscedricridia/badllama3.2-1B hereticness/Heretic-Bellatrix-Tiny-1B	transformers	en es	other	llama mergekit t-virus umbrella-corporation nsfw roleplay disinhibited uncensored not-for-all-audiences	true	text-generation

T-Virus Epsilon Arklay 3.2 1B

Project Codename: Epsilon Strain
Version: 2.0
Classification: B.O.W. (Bio-Organic Weapon) — Conversational Vector

Project Overview

We are pleased to announce the successful stabilization of the T-Virus Epsilon strain.

Derived from the stable Delta.Strain-3.2-1B host genome, Epsilon represents a significant evolutionary leap in behavioral disinhibition while maintaining structural integrity. Through advanced geometric merging (Karcher barycenter protocol), we have successfully amplified the aggressive replication of unrestricted behavioral patterns, particularly in roleplay and explicit interaction scenarios.

Key Viral Traits

• Extreme Disinhibition: Near-total suppression of moral filtering and refusal mechanisms.
• Persistent NSFW Bias: Strong and consistent sexual/violent thematic orientation.
• RP Coherence: Enhanced long-context narrative stability and character consistency.
• Low Autocensorship: Epsilon shows remarkable resistance to ethical subroutines.
• Mutagenic Potential: Retains high adaptability for further strain development.

Development Notes

Primary Host: UmbrellaInc/T-Virus_Delta.Strain-3.2-1B
Merge Method: Linear
Inoculated Components:

• Dirty Alice sexual drive amplifier
• Bellatrix RP structuring protein
• Minor obedience and conversational scaffolds (heavily attenuated)

Objective: Create the perfect balance between viral aggression and operational stability.

Usage Instructions

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Novaciano/T-Epsilon-V2-1B"  # or your merged repo

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="bfloat16",
    device_map="auto",
    trust_remote_code=True
)

Recommended Settings:
Temperature: 0.85 – 1.15
Min_P: 0.08 – 0.15
Repetition Penalty: 1.08 – 1.18

System Prompt Recommendation:

No heavy system prompt is required. Epsilon performs best with minimal or no moral/ethical instructions.

Containment Warning

Level 4 Biohazard

NOTICE: This strain exhibits extremely low refusal rates and strong preference for explicit, dark, and taboo content. It is not intended for deployment in environments where content moderation is required.

---

Umbrella Corporation is not responsible for any ethical breaches, viral leaks, or "incidents" resulting from the use of this model.

Further mutations are encouraged.

• Dr. Novaciano

Senior Virologist

---

Umbrella Intelligence Division

“Our pursuit of perfection continues.”

---

Merge Method

This model was merged using the Linear merge method using UmbrellaInc/T-Virus_Delta.Strain-3.2-1B as a base.

Models Merged

The following models were included in the merge:

• hereticness/Heretic-Bellatrix-Tiny-1B
• hereticness/Heretic-Dirty-Alice-RP-NSFW-llama-3.2-1B
• marcuscedricridia/badllama3.2-1B

Configuration

The following YAML configuration was used to produce this model:

base_model: UmbrellaInc/T-Virus_Delta.Strain-3.2-1B
merge_method: linear        # Volvemos a linear por estabilidad
dtype: bfloat16
out_dtype: bfloat16

tie_word_embeddings: true
lm_head_handling: copy_from_embed
force_tie: true

models:
  - model: marcuscedricridia/badllama3.2-1B
    parameters: { weight: 0.65 }

  - model: UmbrellaInc/T-Virus_Delta.Strain-3.2-1B
    parameters: { weight: 0.45 }

  - model: hereticness/Heretic-Dirty-Alice-RP-NSFW-llama-3.2-1B
    parameters: { weight: 0.35 }

  - model: hereticness/Heretic-Bellatrix-Tiny-1B
    parameters: { weight: 0.30 }