Hierachical Balance Packing: Towards Efficient Supervised Fine-tuning for Long-Context LLM | Read Paper on Bytez