The fine-tuned model "Tanmoy-3.2-1B" is a specialized version of the LLaMA 3.2 1B model, optimized for reflective assistant tasks.
It was trained on the R1 dataset (link:
https://huggingface.co/datasets/ServiceNow-AI/R1-Distill-SFT), which emphasizes thorough,
iterative reasoning and mimics human stream-of-consciousness thinking. The model is designed to explore problems deeply,
express self-doubt, and continuously refine its answers, making it well-suited for complex problem-solving and reflective tasks.
