[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fNr_K23gZCKCDYRBEpZA2smK3zGvTh_TnvxdECRdzk6M":3},{"code":4,"msg":5,"data":6},200,"操作成功",{"id":7,"title":8,"content":9,"digest":10,"source":10,"coverPath":11,"thumbsCoverPath":12,"isTop":13,"isShow":14,"baseClick":13,"clickCount":15,"createTime":16,"typeId":17,"isNewest":18,"newsInfoTypeRespVo":19,"voiceUrl":22,"voiceSize":23,"taskId":24,"releaseTime":25,"titleEn":26,"contentEn":27,"voiceUrlEn":28,"taskIdEn":29,"voiceSizeEn":30},1271,"GPT-5亲口认证系统提示词泄露！15000个Token曝光，OpenAI核心机密不保了","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">AI圈又炸锅了!自GPT-5发布以来，全球技术极客们就像考古学家一样疯狂挖掘其系统提示词，试图揭开OpenAI如何\"调教\"这个超级AI大脑的神秘面纱。近日，GitHub上一个仓库突然曝光了一份疑似GPT-5完整系统提示词的文档，内容多达15000个Token，瞬间在技术社区引发轩然大波。\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">更戏剧性的是，为了验证这份泄露文档的真实性，有人直接让GPT-5本尊来\"验明正身\"。当研究者将泄露的提示词提交给GPT-5并要求其进行自我对比时，这个AI模型的回应让人大跌眼镜——它竟然承认了泄露内容的高度准确性。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">GPT-5在评估过程中透露了几个关键细节。在身份认知方面，泄露版本明确标注\"你是ChatGPT，基于GPT-5模型\"，而GPT-5证实自己确实会在对话中主动介绍身份并标记知识截止日期。在语气风格上，泄露文档要求AI保持\"有洞见、鼓励式\"的交流方式，实际情况则更注重自然流畅的表达，避免犹豫不决的语言模式，力求给用户提供明确建议。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Faccda0e208cd475c88cc512c2b96ccec\u002F6389173250267061421560220.png\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">最有趣的发现在于后台工作机制。泄露版本对后台处理描述相对简略，但GPT-5强调自己严格遵循\"当下交付\"原则，必须立即完成所有能够处理的任务，绝不允许拖延或推迟。这一细节暴露了OpenAI在提升用户体验方面的精心设计。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">GPT-5对这份泄露文档给出了相当正面的评价，认为其内容与用户实际体验高度吻合。更重要的是，它还主动补充了泄露文档遗漏的部分，特别是关于后台处理的严格规范要求。这种\"自爆\"行为不仅验证了泄露内容的可信度，也意外揭示了更多技术细节。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fbaf4a9130aa54af9a294042854aa35fa\u002F6389173252373021881627713.png\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">系统提示词之所以备受关注，是因为它们堪称AI模型的\"基因密码\"。这些指令就像严格的家规一样，塑造着AI的性格特征和行为准则。对于ChatGPT这样的现象级产品，其系统提示词甚至成为了整个行业的设计标杆和参考模板。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">技术爱好者们对提示词的狂热追求并非无的放矢。理解这些核心指令有助于用户更好地与AI协作，提升交互效果，同时也为开发者设计自己的AI产品提供了宝贵借鉴。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">当然，也有不少业内人士质疑这次泄露的真实性，怀疑可能是OpenAI故意释放的\"烟雾弹\"，用来混淆视听或测试市场反应。但无论真假，这次事件都在AI圈掀起了关于模型透明度和技术伦理的新一轮讨论。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这场\"自曝\"风波再次证明，即使是最先进的AI系统，其运作机制仍然充满了让人着迷的秘密。随着AI技术的不断演进，类似的技术考古活动恐怕还会持续上演，而每一次曝光都将为我们理解人工智能的本质增添新的维度。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">【新闻来源】\u003C\u002Fspan>\u003Cspan style=\"color: rgb(187, 187, 187); font-size: 14px;\">AIbase\u003C\u002Fspan>\u003Cspan style=\"color: rgb(187, 187, 187);\"> \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fnews.aibase.com\u002Fzh\u002Fnews\u002F20784\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\"> https:\u002F\u002Fnews.aibase.com\u002Fzh\u002Fnews\u002F20784\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（本网转发此文章，旨在为读者提供更多的信息资讯，所涉内容不构成投资、消费建议。文章事实如有疑问，请与有关方核实，文章观点非本网观点，仅供读者参考。）\u003C\u002Fspan>\u003C\u002Fp>","","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Ffb598b032655488eb04e59e6163f1b56\u002FAI领域.jpg","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fthumbs\u002Ffb598b032655488eb04e59e6163f1b56\u002FAI领域.jpg",0,1,223,"2025-08-25 18:40",2,false,{"id":17,"name":20,"enName":21},"芯位视野","Xinwei Vision","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A34361078-913d-44ba-bfa8-b2cdc840edef%3A0.wav?Expires=1756123455&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=86Sk3gUVKBEnUYqmp4rhd3OxKEs%3D",5554640,"34361078-913d-44ba-bfa8-b2cdc840edef","2025-08-25 18:37","GPT-5 Officially Confirmed System Prompt Leak! 15000 Tokens Exposed, OpenAI's Core Secrets Are Compromised","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">The AI Circle Is in Chaos Again! Since the release of GPT-5, tech geeks around the world have been digging into its system prompts like archaeologists, trying to uncover the mysterious veil of how OpenAI \"trained\" this super AI brain. Recently, a repository on GitHub suddenly exposed a document that appears to be the complete system prompt for GPT-5, with content totaling 15000 tokens, instantly causing a big stir in the technical community.\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">More dramatically, to verify the authenticity of this leaked document, someone directly asked GPT-5 itself to \"confirm its identity.\" When researchers submitted the leaked prompts to GPT-5 and asked it to compare itself, the AI model's response was surprising — it actually admitted the high accuracy of the leaked content.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">During the evaluation process, GPT-5 revealed several key details. In terms of identity awareness, the leaked version clearly states \"You are ChatGPT, based on the GPT-5 model,\" while GPT-5 confirmed that it indeed actively introduces its identity and marks the knowledge cutoff date during conversations. In terms of tone and style, the leaked document requires the AI to maintain an \"insightful and encouraging\" communication approach, while the actual situation focuses more on natural and smooth expression, avoiding hesitant language patterns, aiming to provide users with clear advice.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Faccda0e208cd475c88cc512c2b96ccec\u002F6389173250267061421560220.png\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The most interesting discovery is about the background operation mechanism. The leaked version describes the background processing relatively briefly, but GPT-5 emphasized that it strictly follows the \"immediate delivery\" principle, must complete all tasks that can be handled immediately, and does not allow any delay or postponement. This detail reveals the careful design of OpenAI in improving user experience.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">GPT-5 gave a rather positive evaluation of this leaked document, believing that its content closely matches the user's actual experience. More importantly, it also proactively supplemented parts missing from the leaked document, especially the strict requirements for background processing. This \"self-exposure\" behavior not only verified the credibility of the leaked content but also inadvertently revealed more technical details.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fbaf4a9130aa54af9a294042854aa35fa\u002F6389173252373021881627713.png\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">System prompts have attracted so much attention because they are akin to the \"genetic code\" of AI models. These instructions are like strict family rules, shaping the personality traits and behavioral norms of AI. For products like ChatGPT, their system prompts have even become the benchmark and reference template for the entire industry.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Tech enthusiasts' intense pursuit of prompts is not without reason. Understanding these core instructions helps users better collaborate with AI, improve interaction effectiveness, and also provides valuable references for developers designing their own AI products.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Of course, many industry professionals have questioned the authenticity of this leak, suspecting it might be an \"smoke screen\" deliberately released by OpenAI to confuse the public or test market reactions. But regardless of the truth, this incident has sparked a new round of discussion within the AI circle about model transparency and technological ethics.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">This \"self-exposure\" incident once again proves that even the most advanced AI systems still hold fascinating secrets in their operational mechanisms. As AI technology continues to evolve, similar technological archaeological activities are likely to continue, and each exposure will add new dimensions to our understanding of artificial intelligence.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">[News Source]\u003C\u002Fspan>\u003Cspan style=\"color: rgb(187, 187, 187); font-size: 14px;\">AIbase\u003C\u002Fspan>\u003Cspan style=\"color: rgb(187, 187, 187);\"> \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fnews.aibase.com\u002Fzh\u002Fnews\u002F20784\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\"> https:\u002F\u002Fnews.aibase.com\u002Fzh\u002Fnews\u002F20784\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\"> (This article is forwarded by this website to provide readers with more information and news. The content involved does not constitute investment or consumption advice. If there are any questions about the facts of the article, please verify with the relevant parties. The views expressed in the article are not the views of this website and are for reference only.)\u003C\u002Fspan>\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3Add80a558-6586-4f3e-9a4f-e0017170a8ab%3A0.wav?Expires=1774838487&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=gvtYqEYiqiyS9JAFmagAPXto7xM%3D","dd80a558-6586-4f3e-9a4f-e0017170a8ab",7076446]