Hugging Face Forgotten-Safeword-24B-v4.0: The Safeword Protocol
Forgotten-Safeword-24B-v4.0 is a specialized 24 billion parameter language model developed by ReadyArt, built upon TheDrummer/Cydonia-24B-v2.1. This iteration, dubbed "The Safeword Protocol," is explicitly designed for generating highly explicit and depraved content, pushing the boundaries of what is typically considered acceptable.
Key Characteristics & Training:
- Base Model: Fine-tuned from TheDrummer/Cydonia-24B-v2.1.
- Content Focus: Direct-injected with the "Safeword dataset" to specialize in generating unhinged and depraved scenarios.
- Scenario Re-integration: Includes re-integrated V1.2 scenarios with an improved distribution of extreme kinks.
- Output Balance: Aims for a specific balance between provocative and extremely explicit content.
Intended Use Cases:
This model is intended for highly niche applications requiring the generation of extreme, explicit, and potentially disturbing content. Users should be aware that the model is designed to:
- Generate content that is highly explicit and potentially offensive.
- Void all warranties on one's "soul" due to its content.
- Produce outputs that may lead to questioning the nature of humanity.
Ethical Considerations:
ReadyArt explicitly states that this model will generate content requiring "industrial-grade brain bleach" and advises users to consider the implications of its use. It is recommended for users who specifically seek to explore highly transgressive and explicit themes in creative writing or roleplay, with a strong understanding of the content's nature. For a slightly less extreme version, ReadyArt suggests using "ReadyArt/Forgotten-Abomination-24B-v4.0" and switching to Safeword for more unhinged outputs.