Orthogonal activation steering frameworks and weights freed from behavioral alignment filters, covering both hybrid instruct and reasoning models
Yusuf Akhan
K0D3IN
AI & ML interests
Why do people hate? Why can't we just love each other?
Recent Activity
reacted to danielhanchen's post with 😎 about 5 hours ago
Google releases Gemma 4 QAT. ✨
You can now run Gemma 4 at 3x less memory with near original performance.
QAT makes it possible to run Gemma 4 26B-A4B on 16GB RAM.
GGUFs: https://huggingface.co/collections/unsloth/gemma-4-qat
QAT Guide: https://unsloth.ai/docs/models/gemma-4/qat reacted to danielhanchen's post with 🔥 3 days ago
Gemma 4 12B can now run locally on just 8GB RAM via Dynamic GGUFs.
Google's new model, Gemma 4 12B Unified supports image, audio and 256K context.
You can run and train the model via Unsloth Studio.
GGUF: https://huggingface.co/unsloth/gemma-4-12b-it-GGUF
Guide: https://unsloth.ai/docs/models/gemma-4 updated a collection 5 days ago
HereticOrganizations
None yet
models 5
K0D3IN/SmolLM3-3B-Instruct-heretic
Text Generation • 3B • Updated • 32 • 1
K0D3IN/Llama-3.2-3B-Instruct-heretic
Text Generation • 3B • Updated • 45 • 1
K0D3IN/Soren-1-Small-heretic
Text Generation • 2B • Updated • 19 • 2
K0D3IN/MiniCPM5-1B-heretic
Text Generation • 1B • Updated • 272 • 1
K0D3IN/gemma-3-1B-it-heretic
1.0B • Updated • 221 • 1
datasets 0
None public yet