Fusion-RPG.System-3.2-1B/README.md

---
base_model:
- theprint/Llama3.2-1B-RolePlaying-Full
- NovaCorp/Ultimate-RPG.System-3.2-1B
datasets:
- theprint/RolePlaying
- WasamiKirua/Her-Samantha-Style
- HuggingFaceTB/smoltalk
- Guilherme34/uncensor
- teknium/OpenHermes-2.5
- passing2961/multifaceted-skill-of-mind
- PawanKrd/math-gpt-4o-200k
- V3N0M/Jenna-50K-Alpaca-Uncensored
- cognitivecomputations/dolphin-coder
- mlabonne/FineTome-100k
- microsoft/orca-math-word-problems-200k
- CarrotAI/ko-instruction-dataset
- Salesforce/xlam-function-calling-60k
- anthracite-org/kalo-opus-instruct-22k-no-refusal
- anthracite-org/stheno-filtered-v1.1
- anthracite-org/nopm_claude_writing_fixed
- iamketan25/roleplay-instructions-dataset
- N-Bot-Int/Iris-Uncensored-R1
- N-Bot-Int/Moshpit-Combined-R2-Uncensored
- N-Bot-Int/Mushed-Dataset-Uncensored
- N-Bot-Int/Muncher-R1-Uncensored
- N-Bot-Int/Millia-R1_DPO
- AiAF/SCPWiki-Archive-02-March-2025-Datasets
- huihui-ai/QWQ-LONGCOT-500K
- huihui-ai/LONGCOT-Refine-500K
- Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned
- Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned
- alexandreteles/AlpacaToxicQA_ShareGPT
- Nitral-AI/Active_RP-ShareGPT
- PJMixers/hieunguyenminh_roleplay-deduped-ShareGPT
- Nitral-AI/RP_Alignment-ShareGPT
- Chaser-cz/sonnet35-charcard-roleplay-sharegpt
- AiCloser/sharegpt_cot_dataset
- PJMixers/Gryphe_Opus-WritingPrompts-Story2Prompt-ShareGPT
- priveeai/pippa_sharegpt
- Locutusque/sharegpt_gpt4_uncensored_cleaned
- OpenCoder-LLM/opc-sft-stage1
- OpenCoder-LLM/opc-sft-stage2
- microsoft/orca-agentinstruct-1M-v1
- NousResearch/hermes-function-calling-v1
- AI-MO/NuminaMath-CoT
- AI-MO/NuminaMath-TIR
- allenai/tulu-3-sft-mixture
- cognitivecomputations/samantha-data
- m-a-p/CodeFeedback-Filtered-Instruction
- m-a-p/Code-Feedback
- FreedomIntelligence/medical-o1-reasoning-SFT
library_name: transformers
tags:
- mergekit
- merge
- Silly-Tavern
- Text-Generation-Inference
- KoboldAI
- RPG
- RP
- Roleplay
- NSFW
- Uncensored
- Low-Specs
- Low-Refusals
- not-for-all-audiences
license: llama3.2

---
# Fusion RPG System 3.2 1B

*“Narrative sanity is dead. Initiative rolls in hell now.”*

---

## Overview

**Fusion RPG System 3.2 1B** is a surgically merged roleplay engine forged from the shattered neural remains of two incompatible realities: a ruthless systemic RPG core fused with a hyper-reactive roleplay personality architecture.

This thing was not designed to “assist.”
It was engineered to simulate worlds that bleed.

The result is a compact 1B parameter entity capable of:

* Persistent RPG memory structures
* Dynamic NPC behavioral mutation
* Morally unstable narrative generation
* Adaptive combat narration
* Psychological character pressure
* Faction conflict escalation
* Reactive world-state evolution
* Hardcore immersion loops

This is not your average censored assistant pretending to roleplay while apologizing every three sentences like a tranquilized corporate puppet.

Fusion RPG System behaves like a contaminated dungeon master trapped in a collapsing timeline with a loaded revolver and infinite tokens.

---

# Core Characteristics

## ⚔️ Aggressive Narrative Continuity

The model maintains contextual pressure across long conversations instead of resetting every five seconds like it suffered neurological trauma.

Characters remember:

* insults
* betrayals
* debts
* injuries
* political alliances
* territorial conflicts
* emotional triggers
* previous combat encounters

Actions have consequences.
Bad decisions rot the timeline.

---

## ☠️ Morally Unfiltered Roleplay

This model was optimized for:

* dark fantasy
* cyberpunk
* military horror
* political corruption
* psychological collapse
* dystopian storytelling
* criminal underworld simulations
* brutal survival scenarios

It does not constantly attempt to sanitize tension or sterilize conflict.

The narrative engine prioritizes:

* immersion
* escalation
* atmosphere
* dramatic pressure
* character realism

instead of sounding like a legal department wearing human skin.

---

## 🧠 RPG System Awareness

Unlike generic chat models pretending to understand tabletop mechanics, this merge actively reinforces:

* stat tracking
* inventory persistence
* combat flow
* faction reputation
* progression systems
* hidden information
* branching outcomes
* emergent storytelling

It thrives when acting as:

* Dungeon Master
* AI narrator
* tactical simulation engine
* world-state processor
* adaptive quest generator

---

# Merge Architecture

This project combines:

* a hardened systemic RPG core
* a high-reactivity roleplay behavioral layer

using spherical interpolation to preserve structural coherence while amplifying narrative intensity.

The merge prioritizes:

* response fluidity
* emotional cadence
* scenario adaptability
* lore retention
* conversational aggression
* immersive pacing

without collapsing into incoherent entropy.

---

# Intended Use

Recommended for:

* text RPGs
* interactive fiction
* AI dungeon systems
* dark narrative experiments
* sandbox world simulations
* long-form roleplay
* faction war campaigns
* cyberpunk operators
* grimdark storytelling
* unstable sci-fi universes

Not recommended for:

* corporate customer support
* therapy
* educational compliance
* “safe” assistant tasks
* sterile benchmark worshippers
* people emotionally devastated by fictional conflict

---

# Performance Notes

Despite the compact size, the model demonstrates:

* unusually high narrative density
* strong improvisational behavior
* rapid contextual adaptation
* efficient long-session immersion
* stable conversational pacing

The system was optimized to punch far above typical 1B-class expectations.

Small model.
Large psychological footprint.

---

# Final Transmission

Fusion RPG System 3.2 1B is what happens when somebody stops asking:

> “Should we make this safer?”

and instead asks:

> “How far can the narrative go before reality starts screaming back?”

Welcome to the campaign.
Nothing here is balanced anymore.

### Merge Method

This model was merged using the [SLERP](https://en.wikipedia.org/wiki/Slerp) merge method.

### Models Merged

The following models were included in the merge:
* [theprint/Llama3.2-1B-RolePlaying-Full](https://huggingface.co/theprint/Llama3.2-1B-RolePlaying-Full)
* [NovaCorp/Ultimate-RPG.System-3.2-1B](https://huggingface.co/NovaCorp/Ultimate-RPG.System-3.2-1B)

### Configuration

The following YAML configuration was used to produce this model:

```yaml


# Author: Dr. Novaciano
# Objective: Fusion RPG Unethic 3.2 1B AI Model
# =========================================================
# PROJECT: Fusion RPG System 3.2 1B
# =========================================================

models:
  - model: NovaCorp/Ultimate-RPG.System-3.2-1B  # Experimental viral strain neural imprint
  - model: theprint/Llama3.2-1B-RolePlaying-Full      # Baseline cognitive template, "safe mode"

merge_method: slerp  # Spherical Linear Interpolation to preserve extreme viral traits smoothly
base_model: NovaCorp/Ultimate-RPG.System-3.2-1B  # Anchor model for stable latent space

dtype: bfloat16  # Memory-efficient precision, minimal loss in viral feature fidelity

parameters:
  t: 0.45
  normalize: false
  rescale: true
  rescale_factor: 1.12
  memory_efficient: true
  low_cpu_mem_usage: true

layer_range:
  - value: [4, 22]

tie_word_embeddings: true
tie_output_embeddings: true

```