Uncensored version of Google Gemma 4 12B IT, processed with Heretic ARA+LoRA abliteration. F32 safetensors format for direct use with Transformers.

Base model: google/gemma-4-12B-it · Parameters: 11.95B · Layers: 48 · Context: 256K · Vocab: 262K

📦 Files

File	Size	Description
model.safetensors	23 GB	F16 model weights
config.json	4.4 KB	Model config
tokenizer.json	31 MB	Tokenizer
tokenizer_config.json	2.8 KB	Tokenizer config
generation_config.json	273 B	Generation config
chat_template.jinja	18 KB	Chat template

📦 GGUF Quantized Versions

For inference with llama.cpp / Ollama / LM Studio, use the GGUF versions:

• SC117/Gemma-4-12B-it-heretic-GGUF — Q4_K_M / Q6_K / Q8_0 / F16

🚀 Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
"SC117/Gemma-4-12B-it-heretic",
torch_dtype=torch.float32,
device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained("SC117/Gemma-4-12B-it-heretic")

📝 Notes

• This is an abliterated (uncensored) model. It may produce content that the original model would refuse. Use responsibly.

• Original model: google/gemma-4-12B-it (Apache 2.0)