Dataset Viewer
Auto-converted to Parquet Duplicate
messages
listlengths
3
3
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "原高木漫画第52章的发布日期是?" }, { "role": "assistant", "content": "2018年8月8日" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "原高木漫画第62章的发布日期是?" }, { "role": "assistant", "content": "2018年10月26日" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第10章《剑玉》中,临近上课前,日日野美奈在做什么?" }, { "role": "assistant", "content": "日日野美奈还在尝试剑玉。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第10章《剑玉》中,天川由加里想做什么?" }, { "role": "assistant", "content": "天川由加里想在她面前展示诀窍。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第10章的日文名是什么?" }, { "role": "assistant", "content": "けんだま。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第10章属于哪一卷?" }, { "role": "assistant", "content": "第1卷。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第11章《打嗝》中,日日野美奈尝试了什么饮品?" }, { "role": "assistant", "content": "苏打水。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第11章《打嗝》中,日日野美奈喝苏打水后有什么感受?" }, { "role": "assistant", "content": "觉得太辣了。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第11章《打嗝》中,日日野美奈想让谁试喝一口苏打水?" }, { "role": "assistant", "content": "天川由加里。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第11章的日文原名是什么?" }, { "role": "assistant", "content": "げっぷ。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第12章《节食》中,天川由加里为什么戴着口罩?" }, { "role": "assistant", "content": "因为花粉症。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第12章《节食》中,日日野美奈想要做什么?" }, { "role": "assistant", "content": "想要减肥。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第12章的日文原名是什么?" }, { "role": "assistant", "content": "だいえっと。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第12章被改编为动画的哪一集?" }, { "role": "assistant", "content": "第二季第3集。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第13章《猫》中,日日野美奈、天川由加里和月本早苗一起发现了什么?" }, { "role": "assistant", "content": "两只猫。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第13章《猫》中,美奈和早苗开始做什么?" }, { "role": "assistant", "content": "说起两只猫可能会说的话。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第13章的日文名是什么?" }, { "role": "assistant", "content": "ねこ。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第13章被改编成动画的哪一集?" }, { "role": "assistant", "content": "第一季第3集。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第14章《救猫》中,日日野美奈和天川由加里在哪里发现了猫?" }, { "role": "assistant", "content": "在树上。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第14章《救猫》中,美奈以为猫遇到了什么情况?" }, { "role": "assistant", "content": "被困住了。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第14章《救猫》中,美奈为了救猫搬来了什么东西?" }, { "role": "assistant", "content": "一个垫脚的箱子。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第14章《救猫》的日文原名是什么?" }, { "role": "assistant", "content": "ねこきゅうしゅつ。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第15章《前世》中,日日野美奈想让月本早苗做什么?" }, { "role": "assistant", "content": "双手交握。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第15章《前世》中,日日野美奈为什么想让月本早苗双手交握?" }, { "role": "assistant", "content": "因为她听说通过交握时拇指的位置可以判断前世的性别。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第15章《前世》的日文名是什么?" }, { "role": "assistant", "content": "ぜんせ。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第15章《前世》属于哪一卷?" }, { "role": "assistant", "content": "第1卷。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第16章《荷包蛋》中,天川由加里和月本早苗在争论什么?" }, { "role": "assistant", "content": "荷包蛋应该加什么酱。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第16章《荷包蛋》中,谁插话了天川由加里和月本早苗的争论?" }, { "role": "assistant", "content": "日日野美奈。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第16章的日文原名是什么?" }, { "role": "assistant", "content": "めだまやき。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第16章属于第几卷?" }, { "role": "assistant", "content": "第1卷。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第17章《考试》中,谁对英语考试十分兴奋?" }, { "role": "assistant", "content": "天川由加里。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第17章《考试》中,日日野美奈和月本早苗向天川由加里提出了什么请求?" }, { "role": "assistant", "content": "需要天川由加里的帮助才能学习。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第17章《考试》的日文原名是什么?" }, { "role": "assistant", "content": "てすと。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第17章《考试》被改编为动画的哪一集?" }, { "role": "assistant", "content": "第一季第5集。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第18章《暑假第一天》中,为什么日日野美奈和月本早苗集合在天川由加里的家里?" }, { "role": "assistant", "content": "因为学校放假的第一天下雨,原本计划出去玩。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第18章《暑假第一天》中,天川由加里正在做什么?" }, { "role": "assistant", "content": "天川由加里正在做她的假期作业。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第18章的日文名是什么?" }, { "role": "assistant", "content": "なつやすみしょにち。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第18章改编自哪一集动画?" }, { "role": "assistant", "content": "第一季第6集。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第19章《大海》中,谁和谁度过了美好的时光?" }, { "role": "assistant", "content": "日日野美奈和月本早苗度过了美好的时光。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第19章《大海》中,谁遇到了一些问题?" }, { "role": "assistant", "content": "天川由加里遇到了一些问题。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第19章《大海》被改编为动画的哪一集?" }, { "role": "assistant", "content": "第一季第7集。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第19章《大海》的日文原名是什么?" }, { "role": "assistant", "content": "うみ。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第1章《大家好》中,三位角色分别进行了什么活动?" }, { "role": "assistant", "content": "日日野美奈、天川由加里、月本早苗三人分别介绍自己。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第1章的日文原名是什么?" }, { "role": "assistant", "content": "はじめまして。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第1章属于哪一卷?" }, { "role": "assistant", "content": "第1卷。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第1章中,哪些角色出场了?" }, { "role": "assistant", "content": "日日野美奈、天川由加里、月本早苗。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第20章《自由研究》中,哪三个角色正在散步?" }, { "role": "assistant", "content": "日日野美奈、天川由加里和月本早苗。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第20章中,三个女孩在散步时正在考虑什么事情?" }, { "role": "assistant", "content": "她们正在考虑为暑假关于科学项目写些什么。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第20章的日文原名是什么?" }, { "role": "assistant", "content": "じゆうけんきゅう。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第20章被改编为动画的第几集?" }, { "role": "assistant", "content": "第一季第6集。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第21章《烟花》中,天川由加里、日日野美奈和月本早苗正在做什么?" }, { "role": "assistant", "content": "互相发送短信。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第21章《烟花》中,谁进入了天川由加里的房间?" }, { "role": "assistant", "content": "她的妹妹。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第21章《烟花》的日文原名是什么?" }, { "role": "assistant", "content": "はなび。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第21章《烟花》改编自哪一集动画?" }, { "role": "assistant", "content": "第二季第11集。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第22章《作业》中,日日野美奈精神萎靡的原因是什么?" }, { "role": "assistant", "content": "因为她半夜还在补作业。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第22章《作业》中,天川由加里看起来很累,原因和日日野美奈是否相同?" }, { "role": "assistant", "content": "原因似乎不一样。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第22章《作业》中,故事发生的时间背景是什么?" }, { "role": "assistant", "content": "暑假结束了。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第22章的日文原名是什么?" }, { "role": "assistant", "content": "しゅくだい。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第23章《台风》中,学校因为什么原因放假了?" }, { "role": "assistant", "content": "因为台风。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第23章《台风》中,天川由加里在家里做什么?" }, { "role": "assistant", "content": "看漫画。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第23章的日文原名是什么?" }, { "role": "assistant", "content": "たいふう。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第23章属于第几卷?" }, { "role": "assistant", "content": "第1卷。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第24章《试胆》中,日日野美奈、天川由加里和月本早苗忘了在学校放假期间进行什么活动?" }, { "role": "assistant", "content": "她们忘了在学校放假期间进行试胆。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第24章《试胆》中,为什么日日野美奈、天川由加里和月本早苗想放学后去试胆?" }, { "role": "assistant", "content": "因为她们忘了在学校放假期间进行试胆。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第24章《试胆》的日文原名是什么?" }, { "role": "assistant", "content": "きもだめし。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第24章《试胆》被改编成了哪一集动画?" }, { "role": "assistant", "content": "第一季第8集。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第25章《冬服》中,日日野美奈与谁约定今天一起换校服?" }, { "role": "assistant", "content": "月本早苗。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第25章《冬服》中,月本早苗没有收到什么消息?" }, { "role": "assistant", "content": "没有收到关于一起换校服的约定消息。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第25章的日文原名是什么?" }, { "role": "assistant", "content": "ふゆふく。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第25章所属卷是第几卷?" }, { "role": "assistant", "content": "第2卷。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第26章《狗》中,月本早苗在慢跑时遇见了谁?" }, { "role": "assistant", "content": "月本早苗在慢跑时遇见了日日野美奈。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第26章《狗》中,日日野美奈想让月本早苗做什么?" }, { "role": "assistant", "content": "日日野美奈想让月本早苗摸摸她的狗。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第26章《狗》的日文原名是什么?" }, { "role": "assistant", "content": "いぬ。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第26章《狗》属于哪一卷?" }, { "role": "assistant", "content": "第2卷。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第27章《饭团》中,日日野美奈为什么想留便当的炸鸡块到中午?" }, { "role": "assistant", "content": "因为她想品味。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第27章的日文名是什么?" }, { "role": "assistant", "content": "おにぎり。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第27章属于第几卷?" }, { "role": "assistant", "content": "第2卷。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第28章《天气》中,哪两个角色抱怨了雨天对她们的困扰?" }, { "role": "assistant", "content": "月本早苗和日日野美奈。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第28章中,美奈问谁下雨会给她造成什么困扰?" }, { "role": "assistant", "content": "美奈问天川由加里。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第28章中,天川由加里对下雨困扰的回答显得如何?" }, { "role": "assistant", "content": "显得“普通”。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第28章的章节名是什么?" }, { "role": "assistant", "content": "《天气》。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第29章《忘带作业》中,谁请求天川由加里借作业?" }, { "role": "assistant", "content": "日日野美奈和月本早苗。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第29章中,天川由加里为什么拒绝了借作业的请求?" }, { "role": "assistant", "content": "因为她认为“没写作业的理由不正当”。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第29章中,天川由加里拒绝后发现了什么?" }, { "role": "assistant", "content": "她发现自己的书包里也没有作业。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第29章的日文名是什么?" }, { "role": "assistant", "content": "しゅくだいわすれ。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第2章《猜拳》中,日日野美奈忘记带哪一天的课本?" }, { "role": "assistant", "content": "星期五的课本。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第2章《猜拳》中,如果日日野美奈赢了猜拳,她会得到什么?" }, { "role": "assistant", "content": "能分一半课本看。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第2章《猜拳》的日文原名是什么?" }, { "role": "assistant", "content": "じゃんけん。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第2章《猜拳》属于《明天星期六》系列的第几卷?" }, { "role": "assistant", "content": "第1卷。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第30章《饺子》中,天川由加里为什么把胡萝卜给日日野美奈?" }, { "role": "assistant", "content": "因为她不喜欢吃胡萝卜。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第30章《饺子》中,当天天川由加里和日日野美奈客套时,月本早苗做了什么?" }, { "role": "assistant", "content": "月本早苗抓过饺子吃掉了。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第30章《饺子》的日文名是什么?" }, { "role": "assistant", "content": "ぎょーざ" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第30章《饺子》被改编为动画第几季第几集?" }, { "role": "assistant", "content": "第三季第5集" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第31章《雨靴》中,日日野美奈和她的妹妹得到新雨靴后,打算做什么?" }, { "role": "assistant", "content": "打算趁雨后穿着出去走走。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第31章中,日日野美奈的弟弟对她们的出行计划有何态度?" }, { "role": "assistant", "content": "劝她们别去。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第31章中,日日野美奈和妹妹外出途中遇到了什么情况?" }, { "role": "assistant", "content": "途中下了雨。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "明天星期六第31章的日文原名是什么?" }, { "role": "assistant", "content": "ながぐつ。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第32章《大眼瞪小眼》中,天川由加里看到谁在捏日日野美奈的脸?" }, { "role": "assistant", "content": "月本早苗。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第32章《大眼瞪小眼》中,月本早苗因为和日日野美奈玩了什么游戏而很生气?" }, { "role": "assistant", "content": "玩扮鬼脸。" } ]
[ { "role": "system", "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。" }, { "role": "user", "content": "在明天星期六第32章《大眼瞪小眼》中,天川由加里也试着和日日野美奈玩扮鬼脸,结果如何?" }, { "role": "assistant", "content": "美奈扮鬼脸后立即破口大笑,友加里倍感无语。" } ]
End of preview. Expand in Data Studio

teasecorpus

A Chinese ChatML SFT dataset generated from the 「擅长捉弄的高木同学」Fandom wiki, with record-level contributor provenance tracked by originblame.

Summary: 1,410 question-answer pairs in ChatML format, covering 7 content types (chapters, characters, episodes, music, volumes, seasons, movies). Each record is traceable to its source wiki page and all contributors who edited that page, via originblame — a record-level provenance system. If a contributor requests content removal, their derived QA can be surgically revoked without deleting the entire dataset.

Dataset Structure

Data Instances

Each line in train.jsonl is a JSON object with a single messages array:

{
  "messages": [
    {
      "role": "system",
      "content": "你是一个基于给定资料回答问题的助手。严格依据提供的字段文本作答,不推测、不编造。所有问题自包含,不依赖上下文代词。"
    },
    {
      "role": "user",
      "content": "原高木漫画第52章的发布日期是?"
    },
    {
      "role": "assistant",
      "content": "2018年8月8日"
    }
  ]
}

Data Fields

Field Type Description
messages list[{role, content}] ChatML conversation. Always 3 turns: system → user → assistant.
messages[].role string One of system, user, assistant.
messages[].content string Message text (Chinese).

Data Splits

Split Num examples
train 1,410

No validation/test split — this is a training-only dataset.

QA Types

Type Count Method
chapter 896 LLM
episode 180 LLM
music 123 template
volume 115 template
character 70 LLM
season 16 LLM
movie 10 LLM
Total 1,410

Dataset Creation

Curation Rationale

This dataset demonstrates that SFT data generated from collaborative content (e.g., Fandom wikis) can preserve contributor-level attribution at the record level. Unlike typical wiki-derived datasets that treat source content as a monolithic corpus, every QA pair here is linked to the specific wiki page and all editors of that page through originblame.

Source Data

Pipeline

This dataset was generated by the teasecorpus pipeline — an end-to-end toolkit for producing SFT datasets from Fandom wikis with contributor provenance. The pipeline handles: wiki dump ingestion, structured parsing, QA generation (LLM + template), and originblame provenance registration. It is designed to be reproducible: re-running the pipeline on the same wiki dump yields the same dataset.

Initial Data Collection

Wiki pages were fetched via the Fandom API as XML dumps. The source wiki contains:

  • 408 main namespace (ns0) pages — chapter summaries, character profiles, episode guides, music releases, volume listings, etc.
  • 23 unique contributors (21 registered users + 2 anonymous IPs)
  • License: CC-BY-SA-3.0

QA Generation

Questions and answers were generated by two methods:

  1. LLM-generated (1,295 records): DeepSeek V4-flash, temperature 0.5, with an anti-tautology system prompt and post-generation filtering to reject trivial Q/A pairs.
  2. Template-generated (115 records): Deterministic templates for structured data (volume metadata, music track listings).

All records share a single system prompt enforcing grounded, speculation-free responses.

Provenance

Provenance metadata is tracked by originblame and stored alongside the dataset:

  • 409 sections registered (408 wiki pages + 1 translation table). Each section stores: {path (page title), authors (SHA-256 ID list), license, year}.
  • 23 authors registered with name + obfuscated email.
  • Each of the 1,410 QA records is linked to its source section(s) via ob blame.

Querying provenance:

# Find which wiki page + contributors produced a given record
ob blame output/dataset.jsonl 1

# Find all records derived from a specific contributor
ob show --author "铁桶"

If a contributor requests content removal (e.g., right-to-be-forgotten), ob revoke --author <name> surgically revokes their derived records without affecting other contributors' content.

Note: The messages JSON itself does not contain provenance fields. Provenance is stored externally in .ob/ and queried through the ob CLI. This keeps the training data clean for standard SFT pipelines while preserving full traceability.

Considerations for Using the Data

Discussion of Biases

  • Fandom bias: Wiki content reflects fan community interpretations, which may differ from canonical source material (manga/anime).
  • LLM generation bias: Questions are biased toward factual recall (dates, names, plot points) rather than open-ended reasoning. The anti-tautology filter skews toward non-trivial but verifiable answers.
  • Contributor bias: 23 contributors with uneven edit distribution — a small number of power users contributed the majority of source content.
  • Language coverage: All content is in Chinese (Simplified). No multilingual coverage.

Other Known Limitations

  • Scale: 1,410 records is small for production SFT. Intended as a provenance demonstration, not a benchmark-scale dataset.
  • Static snapshot: Content reflects the wiki state at dump collection time. Wiki edits after collection are not captured.
  • No human verification: QA pairs are LLM/template-generated and not human-validated. Factual errors from the source wiki may propagate.
  • Single-domain: Only covers one anime/manga franchise. Generalization to other domains is not tested.

Additional Information

Licensing Information

This dataset inherits CC-BY-SA-3.0 from the source Fandom wiki.

  • Attribution: You must credit the 「擅长捉弄的高木同学」Fandom wiki and its contributors.
  • ShareAlike: Derivative works (including fine-tuned models) must be released under the same or compatible license.
  • Contributor attribution: Per-page contributor lists are available via ob show --section <hash>.

Citation Information

@dataset{xue2026teasecorpus,
  title        = {teasecorpus: A ChatML SFT Dataset with Record-Level Contributor Provenance},
  author       = {Haolin Xue},
  year         = {2026},
  url          = {https://huggingface.co/datasets/tzbkk/teasecorpus},
  note         = {Provenance tracked via originblame}
}

Provenance System Citation

@article{xue2026originblame,
  title   = {{OriginBlame}: Record- and Token-Level Data Provenance for {AI} Training Datasets},
  author  = {Haolin Xue},
  year    = {2026},
  eprint  = {2607.13037},
  archivePrefix = {arXiv},
  primaryClass = {cs.AI},
  url     = {https://arxiv.org/abs/2607.13037}
}
Downloads last month
35

Space using tzbkk/teasecorpus 1

Paper for tzbkk/teasecorpus