This is a really uncensored version of google/gemma-3-4b-it created with Heretic
https://github.com/p-e-w/heretic

initial Refusals 99/100
-> now 18 Refusals with KL=0.05
18 sounds like a lot, but I added 30 more refusal_markers, so there may be fewer without them.
I also added 17 more bad prompts to the dataset.

Note: This heretic model is highly uncensored; thus use it with extreme caution and care.
you an compare with
https://huggingface.co/MuXodious/gemma-3-4b-it-PaperWitch-heresy


Downloads last month
40
GGUF
Model size
4B params
Architecture
gemma3
Hardware compatibility
Log In to add your hardware

16-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for kalle07/gemma-3-4b-it-heretic_R18_K005-bf16-gguf

Quantized
(209)
this model

Collection including kalle07/gemma-3-4b-it-heretic_R18_K005-bf16-gguf