A fine-tuned version of phi-4 with uncensored responses. This model removes the default safety filters and restrictions, allowing for more open-ended and unrestricted conversations including adult content.

Model Details

Base Model: microsoft/phi-4
Fine-tuning Method: LoRA (Low-Rank Adaptation)
Model Format: Safetensors
License: MIT

Note: This fine-tuned version supports text only. The base model's vision and audio capabilities are not available in this version.

What's Different

Removed refusal behaviors and safety filters
Capable of generating adult content
Responds to prompts the base model would typically decline
Retains the general knowledge and instruction-following capability of Phi-4

ℹ️ Notes on Uncensoring

This model is zero-shot uncensored but not 100% guaranteed. Occasionally the model may still produce a refusal depending on the prompt. If this happens, simply regenerate — it should comply on retry.

Recommended Parameters

These are the parameters used during testing for best results:

Parameter	Value
Temperature	0.85
Top-K	40
Top-P	0.5

GGUF Version

A quantized GGUF version of this model is also available for use with llama.cpp, Ollama, and text-generation-webui: 👉 Phi-4-Uncensored-GGUF

💬 Feedback

If you use this model, I'd love to hear from you! Feel free to:

Leave a comment on this model page
Open a discussion for questions or suggestions
Report any issues or unexpected behaviors

Your feedback helps improve future versions.

⚠️ Disclaimer

This model is intended for research and educational purposes only.

This model is provided as-is, for adults only (18+).

License

This model is licensed under the MIT License.

ccharnkij/Phi-4-Uncensored

Phi-4-Uncensored