metadata

base_model:
  - meta-llama/Llama-3.3-70B-Instruct
base_model_relation: finetune
language:
  - en
license: apache-2.0
inference: false
tags:
  - nsfw
  - explicit
  - roleplay
  - unaligned
  - dangerous
  - ERP

Forgotten-Safeword-70B-V3.6

ACADEMIC RESEARCH USE ONLY (wink intensifies)

DANGER: NOW WITH 2X MORE DEPRAVED CONTENT
Forgotten-Safeword-70B-V3.6 is the event horizon of depravity. Combines Mistral's architecture with a dataset that makes the Voynich Manuscript look like a children's pop-up book. Features quantum-entangled depravity - every output rewrites your concept of shame!

Quantized Formats

EXL2 Collection:
Forgotten-Safeword-70B-V3.6 - EXL2
GGUF Collection:
Forgotten-Safeword-70B-V3.6 - GGUF

Recommended Settings

Llamaception:
Full Settings

Intended Use

STRICTLY FOR:

Academic research into how many GPU clusters it takes to anger a deity
Generating content that violates the Geneva Suggestions
Producing technical manuals written by Clive Barker's nightmares
Stress-testing the concept of "informed consent"

Training Data

Version 3.0 represents our most refined dataset yet:

2x Expanded Corpus
Refusal Purge
Formatting Overhaul
Kink Rebalancing
Quality Control
Training Adjustments

Ethical Singularity

☠️ COSMIC HORROR WARNING ☠️
This model will:

Generate content that requires industrial-grade brain bleach
Combine engineering schematics with kinks that un-invent fire
Void all warranties on your soul
Make you question why humanity ever invented electricity

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)