Hi everyone! Over the past few months I've been working on a research project called ICED (Instruction Compression & Embedding Distillation) , a framework designed to transfer instruction-following and reasoning capabilities from larger language models into much smaller models while preserving as much useful knowledge as possible. The goal is simple: As the first public result of this framework, I've released: 🚀 SLM-FRIDGE-0.5B A 0.5B parameter language model built on the Qwen2....