用「稀有区域指纹」精准追踪大模型血统，一篇学术新方法直接看。
arXiv:2505.12682v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly released under restricted licenses, creating a growi…
揭秘LLM模型蒸馏检测新方法，从学生模型中溯源教师模型，保障AI合规与公平竞争。
arXiv:2607.09692v1 Announce Type: new Abstract: Model distillation -- training on outputs from stronger third-party models -- is widely used to boost …
像细胞日记一样沉淀知识，FROST让每个AI Agent拥有可追溯的记忆族谱
周日慢读：如果细胞会写日记——FROST家族的记忆传承 作者 ：FROST Team 日期 ：2026-07-12 主题 ：轻量科普 | 周日轮换 阅读时间 ：5分钟 一封来自细胞的日记 想象一下，如果你是一个细胞，有一天你突然有了自我意识，会发生什么？ 2026年7月12日 晴 今天是我诞生的第0…
提出用溯源分析检测LLM代理行为失配，防止工具调用偏离用户意图，提升AI安全性。
arXiv:2607.01236v1 Announce Type: cross Abstract: As LLM agents gain increasing access to powerful tools, ensuring that their actions are aligned with…
苹果“Hide My Email”100%可溯源真实邮箱，隐私功能形同虚设，速查iOS更新修复漏洞。
IT之家 7 月 2 日消息，据 404 Media 当地时间 7 月 1 日报道，苹果面向 iCloud+ 订阅用户提供的“Hide My Email”（隐藏我的邮件）功能存在安全漏洞，攻击者可利用该漏洞将匿名生成的转发地址逆向关联至用户的真实 Apple ID 邮箱。 该功能允许付费用户生成以 …
用LLM自主代理自动恢复Issue与Commit的链接，解决代码溯源难题，提升开发协作效率。
arXiv:2508.12232v5 Announce Type: replace-cross Abstract: Issue-to-commit link recovery in software repositories is fundamental to software traceabili…
2026 年让我们再谈 macOS 的画面缩放机制。 查看全文
从非标量信任出发，探讨在代理链中携带类型化退化信息并传播，提升链式系统的可靠性与透明度。
Two posts ago, in the one about agents failing quietly , I handed you a fix for silent degradation: tag a degraded output trust="degraded" , propagate…
Cursor 推出的 AI 生成代码追踪标准格式，让调试与溯源更清晰。
Article URL: https://github.com/cursor/agent-trace Comments URL: https://news.ycombinator.com/item?id=48616614 Points: 1 # Comments: 0
用Ed25519签名AI生成的代码变更，让每次机器代写都有据可查、可追溯。
Article URL: https://github.com/SSX360/matrixscroll Comments URL: https://news.ycombinator.com/item?id=48598284 Points: 2 # Comments: 1
OpenAI公开表态支持欧盟AI内容透明度实践准则，力推溯源标准应对AI生成内容挑战
OpenAI supports the EU Code of Practice on AI content transparency, advancing provenance standards and tools to help people understand AI-generated co…
新型攻击揭示工具型LLM智能体因工件溯源缺口而存在的多步越狱漏洞，安全防御需突破单步文本隔离思维。
arXiv:2606.09084v1 Announce Type: cross Abstract: Tool-using LLM agents interact with the world through actions that persist state in artifacts (e.g.,…
追溯大语言模型中可解释单元的训练数据源头，为理解模型行为提供全新归因方法
arXiv:2601.21996v2 Announce Type: replace-cross Abstract: While Mechanistic Interpretability has identified interpretable circuits in LLMs, their caus…
零Gas费区块链溯源方案，让WooCommerce和Shopify商品供应链透明化，消费者扫码即可验证。
Building a Zero-Gas-Fee Blockchain Traceability System for WooCommerce & Shopify on Polygon In modern agriculture, cosmetics, and luxury goods, tr…
系统梳理LLM Agent中证据追踪与执行溯源方法，为可信AI落地提供关键综述
arXiv:2606.04990v1 Announce Type: cross Abstract: Large language model (LLM)-based agents increasingly solve complex tasks by interacting with externa…
事件溯源设计可扩展分析平台，用户行为事件驱动，状态重建与故障恢复一气呵成。
Designing a scalable event-sourced analytics platform Designing a scalable event-sourced analytics platform In modern data systems, you often need to …
整合 Andrej Karpathy 公开文章、课程与项目，支持 wiki 浏览和智能问答，每次回答都严格引用来源并自动扩展知识图，让提问等于编辑。
I created one simple place to consolidate all of Karpathy's stuff, like his essays, courses and GitHub projects. You can browse it like a normal wiki,…
输入域名就能60秒挖出网站背后的建设者，公开数据一网打尽，技术指纹和注册信息全透明。
I built a little tool that figures out who and what built a site. Just recently updated it. e.g. https://whosbuilt.com/?domain=stripe.com Comments URL…
一个AI系统用8天自我构建完成，每次提交都标记模型作者，人类只需下发任务后喝杯咖啡回来刷新GitHub。
Article URL: https://composio.dev/blog/the-self-improving-ai-system-that-built-itself Comments URL: https://news.ycombinator.com/item?id=48370443 Poin…
这篇论文提出了一个分层框架，系统应对生成式AI带来的真实性债务、合成内容威胁及信任与溯源难题
arXiv:2606.00621v1 Announce Type: cross Abstract: Generative artificial intelligence has fundamentally changed how content is now produced. It has ena…