Phi-4-Uncensored
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.