Falcon3 10b Distilled
A simple, open dataset with Q&A pairs extracted from Falcon3 10b using black box synthetic data generation. Meant for knowledge transfer into smaller edge models.
This dataset was made with a simple automation script written in Python. It wraps around Ollama and runs the 10 billion parameter version of the Falcon3 large language model. The model was prompted to create both high quality prompts and answers, feeding itself with purpose made… See the full description on the dataset page: https://huggingface.co/datasets/uran1um1/falcon3_10b_distilled.
