[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fGqJsdYe7PpY-8twbUGuqDi_E7vr5-LciSYg2k7c8DVc":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":8,"question":12,"related":13,"source":23,"type":86},[],"2026-01-04 13:02:13",301047134,[],{"courseId":9,"courseImg":10,"courseName":11},"53e1d2ef4961cca8eea3e23969ad2cb9","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F03a579384a6dc297c89809b582fcc767.png","默认课程","你知道的大模型的应用都有哪些?大模型的应用你知道都有哪些",[14,25,35,43,48,53,63,72,78,83],{"answer":15,"createTime":5,"id":16,"options":17,"question":22,"source":23,"type":24},[],268086989,[18,19,20,21],"CNN","Transformer","GAN","RNN","大模型主要依赖于哪一类架构作为其核心基础","v1",0,{"answer":26,"createTime":27,"id":28,"options":29,"question":34,"source":23,"type":24},[],"2025-12-16 10:30:33",268086990,[30,31,32,33],"学术论文","知识库网站","私密聊天记录","互联网文本","以下哪项不属于大模型训练数据的来源",{"answer":36,"createTime":27,"id":37,"options":38,"question":41,"source":23,"type":42},[],268086991,[39,40],"正确","错误","大模型只能处理文本数据,不能处理图像或音频",3,{"answer":44,"createTime":27,"id":45,"options":46,"question":47,"source":23,"type":42},[],268086992,[39,40],"预训练是大模型训练中的"通识教育"阶段",{"answer":49,"createTime":27,"id":50,"options":51,"question":52,"source":23,"type":42},[],268086993,[39,40],"大模型的可解释性强,容易追溯决策路径",{"answer":54,"createTime":5,"id":55,"options":56,"question":61,"source":23,"type":62},[],301047130,[57,58,59,60],"安全与伦理风险","可解释性强","计算成本高","数据依赖性高","大模型的局限性包括哪些",1,{"answer":64,"createTime":5,"id":65,"options":66,"question":71,"source":23,"type":62},[],301047131,[67,68,69,70],"仅适用于图像任务","依赖GPU\u002FTPU","长序列处理能力强","并行计算","Transformer模型的核心优势包括哪些",{"answer":73,"createTime":5,"id":74,"options":75,"question":76,"source":23,"type":77},[],301047132,[],"大模型训练分为两个阶段:预训练和________",2,{"answer":79,"createTime":5,"id":80,"options":81,"question":82,"source":23,"type":77},[],301047133,[],"Transformer模型通过________机制动态捕捉序列中单词之间的关联性,模拟人类理解语言时的注意力分配方式",{"answer":84,"createTime":5,"id":6,"options":85,"question":12,"source":23,"type":86},[],[],4]