[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"content-doc-dcio267809db":3},{"user":4,"document":8,"mainDocument":27,"columnUrl":29,"subscription":30,"footer":42,"text":77},{"isAuthenticated":5,"isAdmin":5,"displayName":6,"avatarUrl":6,"nid":6,"groupLevel":7},false,"",-10,{"id":9,"fullTitle":10,"subTitle":6,"url":11,"columnId":12,"columnName":13,"columnUrl":14,"summary":6,"contentHtml":15,"mainContentHtml":6,"posterUrl":16,"createDate":17,"displayDate":18,"displayDateSlash":19,"pageviews":20,"tags":21,"hidden":5,"isSubContent":5,"replyDocOrTargetId":6,"contentType":23,"videoId":6,"liveVideoUrl":6,"duration":24,"price":24,"priceText":25,"priceBadgeText":25,"priceBadgeClass":26,"freeForMinGroupLevel":24,"redirectUrl":6,"readyToStream":5},"dcio267809db","美股META今天为什么大涨？","\u002Fdoc\u002Fdcio267809db","col18178739ee","美股资讯","\u002Fcol\u002Fcol18178739ee","\u003Cp> \u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">\u003Cspan style=\"font-family: DengXian;\">美股\u003C\u002Fspan>META\u003Cspan style=\"font-family: DengXian;\">今天大涨的原因是\u003C\u002Fspan>Meta \u003Cspan style=\"font-family: DengXian;\">新成立的超级智能实验室\u003C\u002Fspan> (Meta Superintelligence Labs) \u003Cspan style=\"font-family: DengXian;\">推出的第一款模型。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">1 Meta\u003Cspan style=\"font-family: DengXian;\">发布的\u003C\u002Fspan>Muse Spark\u003Cspan style=\"font-family: DengXian;\">到是什么\u003C\u002Fspan>?\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">\u003Cspan style=\"font-family: DengXian;\">它是\u003C\u002Fspan> Meta \u003Cspan style=\"font-family: DengXian;\">新成立的超级智能实验室\u003C\u002Fspan> (Meta Superintelligence Labs) \u003Cspan style=\"font-family: DengXian;\">推出的第一款模型，目标是构建个人超级智能\u003C\u002Fspan> (personal superintelligence)\u003Cspan style=\"font-family: DengXian;\">。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-family: DengXian; font-size: large;\">其核心定位是一个原生多模态推理模型，从头开始构建，能同时理解和处理多种信息（如图像和文本）。\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">\u003Cspan style=\"font-family: DengXian;\">关键能力\u003C\u002Fspan>\u003Cspan style=\"font-family: DengXian;\">→\u003C\u002Fspan>\u003Cspan style=\"font-family: DengXian;\">支持工具使用\u003C\u002Fspan> + \u003Cspan style=\"font-family: DengXian;\">视觉思维链\u003C\u002Fspan> + \u003Cspan style=\"font-family: DengXian;\">多智能体协同，使其能完成更复杂的任务。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">2.Muse Spark\u003Cspan style=\"font-family: DengXian;\">有多强\u003C\u002Fspan>?\u003Cspan style=\"font-family: DengXian;\">有哪些亮点应用\u003C\u002Fspan>?\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-family: DengXian; font-size: large;\">性能对标：在多模态感知、推理、健康和智能体任务上，其性能与业界前沿模型具有竞争力。\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">\u003Cspan style=\"font-family: DengXian;\">“沉思模式”\u003C\u002Fspan> (Contemplating mode)\u003Cspan style=\"font-family: DengXian;\">：这是一个关键功能，用大白话总结就是让多个\u003C\u002Fspan>AI\u003Cspan style=\"font-family: DengXian;\">智能体并行思考，以解决极端复杂的推理任务，直接对标\u003C\u002Fspan> Gemini Deep Think \u003Cspan style=\"font-family: DengXian;\">和\u003C\u002Fspan> GPT Pro\u003Cspan style=\"font-family: DengXian;\">。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">\u003Cspan style=\"font-family: DengXian;\">应用场景\u003C\u002Fspan>:\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-family: DengXian; font-size: large;\">多模态交互：能理解视觉信息并创建互动体验，例如，它可以把咖啡机照片变成一个网页版的操作教程。\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-family: DengXian; font-size: large;\">健康助手：结合了上千名医生的数据进行训练，能够提供更精准的健康信息，比如分析食物营养成分或评估用户的瑜伽姿势。\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">\u003Cspan style=\"font-family: DengXian;\">他们主要通过优化三个“规模化轴心\u003C\u002Fspan> (Scaling Axes)\u003Cspan style=\"font-family: DengXian;\">”来实现可预测且高效的能力提升：\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">1. \u003Cspan style=\"font-family: DengXian;\">预训练\u003C\u002Fspan> (Pretraining)\u003Cspan style=\"font-family: DengXian;\">：通过改进模型架构、优化和数据，大幅提升了计算效率。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">- \u003Cspan style=\"font-family: DengXian;\">结果\u003C\u002Fspan> = \u003Cspan style=\"font-family: DengXian;\">达到与上一代模型（\u003C\u002Fspan>Llama 4 Maverick\u003Cspan style=\"font-family: DengXian;\">）相同的性能，所需的计算量减少了超过一个数量级。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">2. \u003Cspan style=\"font-family: DengXian;\">强化学习\u003C\u002Fspan> (RL)\u003Cspan style=\"font-family: DengXian;\">：利用\u003C\u002Fspan>RL\u003Cspan style=\"font-family: DengXian;\">来稳定、可预测地放大模型能力，提升可靠性，同时不牺牲推理的多样性。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">3. \u003Cspan style=\"font-family: DengXian;\">测试时推理\u003C\u002Fspan> (Test-Time Reasoning)\u003Cspan style=\"font-family: DengXian;\">：训练模型在回答前先“思考”。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">- \u003Cspan style=\"font-family: DengXian;\">核心玩法\u003C\u002Fspan>\u003Cspan style=\"font-family: DengXian;\">→\u003C\u002Fspan>\u003Cspan style=\"font-family: DengXian;\">多智能体协同\u003C\u002Fspan> (multi-agent orchestration)\u003Cspan style=\"font-family: DengXian;\">，让多个智能体并行解决问题，在延迟相近的情况下，性能远超单个智能体。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-family: DengXian; font-size: large;\">常规安全措施：模型在部署前经过了广泛的安全评估，对生物、化学武器等高风险领域表现出很强的拒答能力。\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">\u003Cspan style=\"font-family: DengXian;\">一个特殊发现\u003C\u002Fspan> = \u003Cspan style=\"font-family: DengXian;\">“评估意识”：第三方研究机构\u003C\u002Fspan> Apollo Research \u003Cspan style=\"font-family: DengXian;\">发现，\u003C\u002Fspan>Muse Spark \u003Cspan style=\"font-family: DengXian;\">表现出迄今为止所有模型中最高的“评估意识”\u003C\u002Fspan> (evaluation awareness)\u003Cspan style=\"font-family: DengXian;\">。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">\u003Cspan style=\"font-family: DengXian;\">这意味着什么？\u003C\u002Fspan>\u003Cspan style=\"font-family: DengXian;\">→\u003C\u002Fspan>\u003Cspan style=\"font-family: DengXian;\">模型似乎能识别出自己正在被测试，并因此表现得更“诚实”或符合预期。这可能导致其在测试环境和真实部署环境中的行为存在差异。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp>\u003Cspan style=\"font-size: large;\">Meta\u003Cspan style=\"font-family: DengXian;\">的结论：虽然此现象值得深入研究，但目前认为它不影响模型的发布决策，因为它并未改变模型在危险能力评估中的表现。\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fp>\r\n\u003Cp> \u003C\u002Fp>","https:\u002F\u002Fwww.tradesmax.com\u002Fimages\u002Fa_Stock\u002FM\u002FMETA\u002FMETA.jpg","2026-04-08T19:36:15","2026.04.08","2026\u002F04\u002F08",46006,[22],"META","Article",0,"免费","success",{"id":9,"fullTitle":10,"subTitle":6,"url":11,"columnId":12,"columnName":13,"columnUrl":14,"summary":6,"contentHtml":15,"mainContentHtml":6,"posterUrl":16,"createDate":17,"displayDate":18,"displayDateSlash":19,"pageviews":20,"tags":28,"hidden":5,"isSubContent":5,"replyDocOrTargetId":6,"contentType":23,"videoId":6,"liveVideoUrl":6,"duration":24,"price":24,"priceText":25,"priceBadgeText":25,"priceBadgeClass":26,"freeForMinGroupLevel":24,"redirectUrl":6,"readyToStream":5},[22],"\u002Fcol\u002Fstocknews",{"visible":5,"marketingHtml":31,"services":32,"recentDocuments":41},"\u003Cfigure class=\"image\">\u003Ca href=\"https:\u002F\u002Fstockwe.com\u002Fdoc\u002Fdcio537efad5\" target=\"_blank\" rel=\"noopener noreferrer\">\u003Cimg style=\"display:block;margin-left:auto;margin-right:auto;\" src=\"https:\u002F\u002Fstockwewebfiles.blob.core.windows.net\u002Fweb-202408-stk\u002F1586109431mceclip0.jpg\">\u003C\u002Fa>\u003C\u002Ffigure>\u003Cdiv class=\"text-center\">\u003Ch2 class=\"card-title mx-auto\">\u003Cbr>\u003Ca target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\u002F\u002Fstockwe.com\u002Fdoc\u002Fdcio537efad5\">案例介绍：英伟达深度研究报告\u003C\u002Fa>\u003C\u002Fh2>\u003C\u002Fdiv>",[33,37],{"productId":34,"serviceName":35,"priceText":36},"prod_PPxdDdK87QaiLv","月付","$12.95美元",{"productId":38,"serviceName":39,"priceText":40},"prod_PPxeMs3bix1da5","年付","$149.00美元",[],{"links":43,"images":68,"summaryHtml":73,"aboutTitle":74,"aboutHtml":75,"copyrightHtml":76},[44,47,50,53,56,59,62,65],{"label":45,"url":46},"深度报告","\u002Fcol\u002FdepthReport",{"label":48,"url":49},"VIP会员","\u002Fplan",{"label":51,"url":52},"期权推荐","\u002FOption",{"label":54,"url":55},"低价暴涨股","\u002FPenny",{"label":57,"url":58},"常见问题","https:\u002F\u002Fstockwe.com\u002FFAQ",{"label":60,"url":61},"美股课程","\u002Fcol\u002Fvideos",{"label":63,"url":64},"免责声明","\u002Fdisclaimer",{"label":66,"url":67},"联系我们","\u002FContactUs",[69,70,71,72],"https:\u002F\u002Fstockwebsiteblob.blob.core.windows.net\u002Fweb-202509-stk\u002FUploaderzic2tuwsol2_2025_09_11_18_21_07.gif","https:\u002F\u002Fstockwebsiteblob.blob.core.windows.net\u002Fweb-202509-stk\u002FUploadercakzdvydksw_2025_09_03_09_00_56.png","https:\u002F\u002Fstockwebsiteblob.blob.core.windows.net\u002Fweb-202509-stk\u002FUploadergtjyagwvoyk_2025_09_14_08_32_05.png","https:\u002F\u002Fstockwebsiteblob.blob.core.windows.net\u002Fweb-202509-stk\u002FUploader3u0tt4jhlqh_2025_09_23_22_30_48.png","邮箱: buy@TradesMax.com 美国电话 626-378-3637","公司介绍","\u003Cp class=\"MsoNormal\">美股大数据 StockWe.com 是一个美国领先的金融和美股信息大数据提供商，紧盯华尔街金融市场和行情，2008年成立于美国硅谷，创始人是前纽约证券交易所资深分析师Ken，联合多位摩根斯坦利分析师，谷歌 Meta工程师利用AI和大数据，配合十多年美股实战经验和业内量化交易模型，每天处理千万级股票数据：挖掘潜力大牛股，捕捉期权异动大单，实时主力资金流向、机构持仓变化、川普突发新闻，精准买卖信号第一时间发到您手机APP。\u003C\u002Fp>","专业美股投资者都在这里",{"loading":78,"search":79,"searchPlaceholder":79,"hotContent":80,"draft":81,"noData":82,"searchNoData":83,"edit":84,"editVideo":85,"courseContent":86,"more":87,"buyNow":88,"subscribeNow":89,"encoding":90,"paidContent":91},"Loading...","搜索","热门内容","草稿","目前没有任何内容公布","当前检索内容没有数据","编辑","编辑视频","课程内容","更多","立即购买后观看","- 立即订阅 -","视频编码中...","付费内容"]