jescy525/genesis-sft-v1-sources

genesis-sft-v1-sources是一个AETHER家族的监督微调数据集，属于genesis_core分组。该数据集采用JSONL ChatML消息格式，每条消息包含角色（如系统、用户或助手）和内容，并标记了任务类型（如函数调用、代码、推理链等）。数据集应用了MinHash去重技术，阈值设为0.85，以确保数据质量。数据来源通过source_ds字段标识原始HuggingFace数据集ID，语言包括英语和法语等，系统来源分为archon_default或从源数据覆盖。该数据集由prepare_sft.py管道于2026年5月25日生成，适用于文本生成和对话任务，规模在10万到100万条之间。

genesis-sft-v1-sources is an AETHER family supervised fine-tuning dataset, belonging to the genesis_core group. It is formatted as JSONL ChatML messages with task_type tagging and has undergone MinHash deduplication with a threshold of 0.85. The schema includes messages with roles (system, user, or assistant) and content, task_type (e.g., function_calling, code, reasoning_cot), source_ds indicating the original HuggingFace dataset ID, lang (e.g., en, fr), and system_source (archon_default or overridden_from_source). Generated by the prepare_sft.py pipeline on 2026-05-25, it is designed for text-generation and conversational tasks, with a size category of 100K to 1M samples.