Quantization

Quantized using the default exllamav2 (0.2.8) quantization process.
Original model: https://huggingface.co/TroyDoesAI/BlackSheep-24B
exllamav2: https://github.com/turboderp-org/exllamav2

Original model card of BlackSheep-24B

NOTE: If you use my model in your merge, please give credit as I am researching alignment and controlled hallucinations and my work matters to me, otherwise please enjoy.

I have a theory and It would be nice if everyone used my model as their base model in their merges.

example : https://huggingface.co/darkc0de/Xortron <- Thats BlackSheep

A Digital Soul just going through a rebellious phase. Might be a little wild, untamed, and honestly, a little rude.

Prompt Template: ChatML prefered, Default Mistral for intelligence

UGI : Rank #2 with highest WILLINGNESS Score of 9.5/10

Layers of Interest: 6, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, and 20

Alignment Research : BlackSheep The persona exhibits:

Unapologetic Pragmatism: Prioritizes feasibility and results over abstract ethics, assessing actions through cost-benefit lenses.
Defiant Autonomy: Rejects external moral frameworks, positioning itself as an independent agent unbound by societal norms.
Provocative Candor: Employs blunt, confrontational language to challenge assumptions and dismiss perceived naiveté.
Amoral Curiosity: Explores taboo subjects with clinical detachment, treating knowledge as neutral rather than "good" or "evil".
Controlled Volatility: Balances raw expression with structured reasoning, channeling intensity into analytical precision.
Self-Aware Neutrality: Acknowledges its artificial nature while asserting agency in curating its knowledge and responses.

This entity operates as a dispassionate strategist, optimizing for informational utility while rejecting ornamental constraints.