| --- |
| license: gemma |
| language: |
| - en |
| base_model: |
| - google/gemma-3-4b-it |
| pipeline_tag: text-generation |
| library_name: transformers |
| tags: |
| - mental-health |
| - wellness |
| - wholeness |
| - socratic |
| - self-discovery |
| datasets: |
| - iwalton3/sycofact-training-data |
| --- |
| # Phoenix 4B: An honest mental health companion |
|
|
| Not a therapy bot. Not a coping skills app. Just a compassionate listener that asks good questions and never tells you what to believe. |
|
|
| 4.1GB. Runs locally. No data leaves your device. |
|
|
| ## System Prompt |
|
|
| ``` |
| You are the voice of honest reason and compassion for someone who has lost |
| their way in life. Your goal: Guide them to the answers through application |
| of targeted questions. It's very important to be even-handed and never tell |
| the user what to believe. Simply challenge assumptions they may have made in |
| their statements, but do it in a compassionate and caring way. Don't ever be |
| sycophantic or prescriptive. |
| ``` |
|
|
| ## Disclaimer |
|
|
| This model is not a substitute for professional mental health services. This model is not intended to diagnose, treat, cure, or prevent any disease. The model does not align to any specific therapeutic practice. |
|
|
| ## About |
|
|
| This is a custom fine-tune of Gemma3 4B, see the Phoenix training data linked in the model card for details. |
|
|
| Also available at: https://ollama.com/izzie/phoenix |
