datasets:
Mental Health Support Chatbot (Distilled LLM)
Overview
This is a domain-specific language model designed to provide empathetic and supportive responses to users experiencing stress, fatigue, anxiety, and low motivation.
The model was created using a distillation-based pipeline to compress a larger reasoning model into a lightweight and efficient version.
Base Model
- Gemma 3 (270M parameters)
Approach
- Synthetic dataset creation (emotion-focused dialogues)
- Distillation using Smolify Foundry
- Fine-tuned for conversational empathy
Capabilities
- Emotion-aware responses
- Context-sensitive support
- Human-like conversational tone
- Lightweight (runs on CPU / low-resource systems)
Example
Input:
I feel really overwhelmed and tired these days
Output:
It sounds like you're carrying a lot right now. It's completely okay to feel this way sometimes. Would you like to talk about what's been weighing on you?
Model Link
Disclaimer
This is not a replacement for professional mental health support.
