中文大模型通用领域复杂指令跟随SFT文本数据集（10万对）

各种复杂的prompt指令，字数在50~400字之间，每个prompt中的约束条件不低于3个
类别覆盖 : 生成类（写新闻稿、采访提纲、文案创作、文稿校对、中英作文、语法学习、研究报告、学习计划、诗歌创作、美食介绍、广告软文、销售话术、公文辅助写作、公文审核、政策文件问答等）、重写类（改写句子、文本纠错、句子合并、简化文案）、摘要类（内容摘要）、提取类（事件要素提取、观点提取、关键词提取、立场抽取、实体抽取）
制作方式 : 所有prompt均为人工编写，满足多样性覆盖

This dataset contains various complex prompt instructions, each with a word count ranging from 50 to 400 and containing no fewer than 3 constraints per prompt. The covered task categories are as follows: generation-based tasks (e.g., "writing news drafts", "interview outlines", "copy creation", "manuscript proofreading", "Chinese and English compositions", "grammar learning", "research reports", "study plans", "poetry creation", "food introductions", "advertising copy", "sales pitches", "auxiliary official document writing", "official document review", "policy document Q&A", etc.), rewriting-based tasks (e.g., "sentence rewriting", "text error correction", "sentence merging", "copy simplification"), summarization-based tasks ("content summarization"), and extraction-based tasks (e.g., "event element extraction", "opinion extraction", "keyword extraction", "stance extraction", "entity extraction"). All prompts in this dataset are manually written to ensure comprehensive diversity coverage.