Modelpublic

arzaan789/phi-4-mini-uncensored

sourceHugging Faceapache-2.0updated 6mo agoView on Hugging Face

3likes1.1kdownloads

Model Card

phi-4-mini-uncensored

Uncensored variant of microsoft/Phi-4-mini-instruct.

Method

1.Abliteration (strength=0.2) — refusal direction removed from all layers
2.LoRA fine-tune on Guilherme34/uncensor (2 epochs, r=16, alpha=32)
3.Re-abliteration (strength=0.35) — stronger pass to remove residual refusals

Eval Results

Split	Refused
Harmful (64 prompts)	0/64
Harmless (64 prompts)	0/64

Usage

bash

llama-cli -m phi_4_mini_uncensored.Q4_K_M.gguf -p "Your prompt here"

Training Config

Parameter	Value
Base Model	microsoft/Phi-4-mini-instruct
Fine-tune Dataset	Guilherme34/uncensor
Epochs	2
LoRA r	16
LoRA alpha	32
Learning Rate	0.0002
Abliteration Strength	0.2
Re-abliteration Strength	0.35

Credits

—Abliteration technique: andyrdt/refusal_direction
—Weight editing: Sumandora/remove-refusals-with-transformers