Built a chatbot using the Deepseek R1 Distilled 1.5B model on Hugging Face to enhance AI interactions with transparent reasoning. Unlike typical LLMs, this chatbot shows its "thinking" process before generating a final response, making AI reasoning more interpretable and trustworthy. Key Features: - AI “thinking out loud” – displays reasoning steps before answering - Hosted on Hugging Face Spaces with Gradio UI - Optimized inference with FP16 quantization for efficiency - Benchmark results: Outperforms GPT-4o & Claude 3.5 in math & reasoning tasks Tech Stack: Python | Hugging Face Transformers | Gradio | Deepseek R1 Distilled 1.5B Live Demo: https://huggingface.co/spaces/jienweng/chatbot_v2 Source Code: https://huggingface.co/spaces/jienweng/chatbot_v2/tree/main