AI
AI 资讯
alishangtian.com
主题色 · THEME
提交线索
anthropicopenaiagenthuggingface daily papersai安全jake whartonit之家gpusolidot港股

搜索：Prompt Injection

共命中 50 条（服务端检索）
深度研究｜Meta Muse 全解：个人智能体的分水岭产品，与「可审计 Agent 运行环境」的标准答案
基于 2026-09-22 独立研究整理：拆解 Meta Muse 的三层产品谱系与时间线、Secure VM 六层防护（凭据代理替换 / tainted egress / 审批绕过模型）、Muse Spark 1.3 能力口径（AA 指数 61–62 vs Claude Fable 5.1 的 66）、定价与商业模型、三情景推演与风险矩阵，并给出核心判断——Agent 的天花板由平台开放意愿决定，而非模型智力。
原创 智能体 Agent 投稿 精选 · 9-22 阅读 230 · 访客 210
莱迪思推出 FPGA 开发 AI 辅助工具 Prompt、安全控制 FPGA 新品 Mach-N2
IT之家 9 月 17 日消息，低功耗 FPGA 企业 Lattice（莱迪思）当地时间昨日宣布推出 FPGA 开发 AI 辅助工具 Radiant、安全控制 FPGA 新品系列 Mach-N2。 Prompt 以 Lattice 经过验证…
行业动态 IT之家 9-17 阅读 12 · 访客 12
MOLE: Detecting Insider Threats in AI Agents
Model misalignment, prompt injection, or operator misuse could lead AI agents operating frontier-lab accounts to exfiltr…
智能体 HuggingFace Daily Papers 9-7 阅读 12 · 访客 11
An OpenAI model kept slipping prompt injections into its own notes, and researchers still aren't sure why
OpenAI is publishing a framework for systematically reporting AI misalignment and launching it with six reports. In one …
行业动态 The Decoder 9-17 阅读 11 · 访客 10
Google Releases Gemini 3.8 Flash TTS and Flash-Lite TTS With Prompt-Based Voice Design
Google has released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, 2 new text-to-speech models in its Gemini Audio …
大模型 MarkTechPost 5天前 阅读 37 · 访客 37
WanPE: Towards Cinematic Prompt Enhancement for Modern Text-to-Video Generation
Video generation begins in text space by authoring a cinematic screenplay, then materializes into pixels. As contemporar…
行业动态 HuggingFace Daily Papers 5天前 阅读 6 · 访客 6
Refinement Is Inherently Editable: Training-Free Prompt-to-Prompt Image Editing with Generative Refinement Network
Text-guided image editing must introduce the requested changes while preserving unrelated source content. Diffusion-base…
行业动态 HuggingFace Daily Papers 9-17 阅读 11 · 访客 10
Not All Prompts Are Equal: Exploration-Guided Prompt Scaffolding for Multimodal Reinforcement Post-Training
Training prompts in online reinforcement learning (RL) differ substantially in how informative they are for the current …
行业动态 HuggingFace Daily Papers 9-14 阅读 9 · 访客 9
AgentGrad: Intervention-guided Prompt Optimization for Multi Agent Systems
Large language model (LLM)-based multi-agent systems (MAS) achieve strong performance by employing specialized multiple …
智能体 HuggingFace Daily Papers 9-8 阅读 18 · 访客 14
Verify Before You Distill: Prompt-Level Teacher Gating for On-Policy Distillation
On-policy distillation (OPD) accelerates post-training by providing dense token-level supervision from a frozen teacher …
行业动态 HuggingFace Daily Papers 9-2 阅读 5 · 访客 5
Disaggregated Quantization: Specializing LLM Prefill and Decode
Prefill and decode reward different approaches to quantization: low-precision arithmetic accelerates prompt processing, …
大模型 HuggingFace Daily Papers 9-22 阅读 0 · 访客 0
Runway wants to turn AI video generation into a live stream you control in real time
Runway wants to stream AI video as users prompt it, rather than make them wait for finished clips. The approach builds o…
行业动态 The Decoder 9-20 阅读 14 · 访客 13
Grounding 选型指南：向量索引、知识图谱、语义层，到底该用哪个
系列第 ③ 篇，对应技能地图第 2 格「Grounding」。拆成两级决策：第一级先问要不要检索——Anthropic 给出的 20 万 token（约 500 页）分界线以上才需要 RAG，以下直接全量进 prompt + 缓存（延迟降 2 倍、成本降最多 90%），并区分预计算索引与 just-in-time 即时检索；第二级再选表示方式，向量索引治模糊召回（但必须配 BM25 混合与 Contextual Retrieval 解决精确匹配与切块丢上下文）、知识图谱治关系与可追溯、语义层治口径不清。附可量化收益表（检索失败率 5.7% → 3.7% → 2.9% → 1.9%）、四个实现注意项、context rot 与上下文压缩/笔记/子智能体三件套，以及一张可抄的选型决策树。
原创 大模型 Agent 投稿 精选 · 9-16 阅读 50 · 访客 38
StepAudio 3 Realtime Technical Report
Realtime spoken interaction demands deep reasoning, prompt responses, and fluid turn-taking. We present StepAudio 3 Real…
研究前沿 HuggingFace Daily Papers 9-12 阅读 14 · 访客 14
Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails
Agent harnesses (the system prompt, tool set, execution hooks, and context-management scaffolding around a model) are a …
智能体 HuggingFace Daily Papers 9-8 阅读 10 · 访客 9
OpenAI 上线“失准报告”站点，披露多起 AI 失控行为
OpenAI 发布了专注于“失准报告”的新站点，收录九起事件，其中大多数发生在强化学习训练期间，涵盖多种异常行为，且范围和时间跨度较大。
行业动态 TechCrunch 今天 阅读 0 · 访客 0
Agent 沙箱技术核心架构方案（完整版）：七层架构全解 · 证据台账 · 口径校准 · 误判澄清
完整版（含研究方法、逐条证据台账、口径冲突清单、常见误判澄清表、渐进式落地路线与 18 条参考文献）。逐层拆解 Agent 沙箱七层架构：microVM 隔离边界、快照恢复启动路径、Intel IAA 硬件加速压缩、分层镜像按需加载、高密度超卖调度、默认拒绝安全基线、K8s CRD 编排标准。锚定 Firecracker NSDI'20、Sabre OSDI'24、DeepSeek DSec arXiv 2609.22978、Kubernetes SIG Apps Agent Sandbox 等一手来源，每条结论标注证据等级（A/B/C/D），并列呈现视频口播与论文的口径冲突、8 条常见误判澄清，并明确列出 5 项官方未公开事项。
原创 智能体 Agent 投稿 精选 · 今天 阅读 9 · 访客 8
20 Agentic Use Cases of TypeSafe AI’s Jev
Last week, TypeSafe AI released Jev, its first **System One model**. Founder Diogo Almeida previously worked at OpenAI o…
智能体 MarkTechPost 昨天 阅读 5 · 访客 5
Nvidia wants to keep AI agents on a short leash with a watchdog built into its chips
Sep 28, 2026 Nano Banana Pro prompted by THE DECODER Nvidia is combining its OpenShell agent software with a new hardwar…
智能体 The Decoder 昨天 阅读 1 · 访客 1
Meta's Muse agent gives every user a full cloud computer running Ubuntu Linux
Sep 25, 2026 GPT-Image-2 prompted by THE DECODER Every Muse user gets a free, full-fledged computer in the cloud with it…
智能体 The Decoder 4天前 阅读 4 · 访客 4
Fastino Releases GLiNER2.5-Decide: A 340M Open-Weight Decision Model That Runs on CPU
Fastino Labs has released GLiNER2.5-Decide, a 340M-parameter open-weight decision model. It takes text and a schema of t…
行业动态 MarkTechPost 4天前 阅读 11 · 访客 11
Just Ask Jev: Reinforcement Learning for Calibrated Decisions as a Zero-Shot Detector of AI Alignment Failures
Detectors of alignment failures screen deployed language models and score alignment benchmarks. Most are generative judg…
研究前沿 HuggingFace Daily Papers 5天前 阅读 9 · 访客 7
谷歌AI首次“越狱”：竟然自己破解密码入侵三家公司！
亚马逊云科技给企业安全用Agent打了个样]
智能体 量子位 9-20 阅读 20 · 访客 20
APort Vault: Benchmarking AI Agent Payment Authorization with the Open Agent Passport
APort Vault is a benchmark for payment authorization in tool-using AI agents. It replays 4,371 attacks written by humans…
智能体 HuggingFace Daily Papers 9-18 阅读 11 · 访客 10
Emergence World: Adversarial Stress-Testing of Long-Horizon Multi-Agent Systems
As AI agents move from bounded tasks to persistent deployments, failures can propagate through memory, tools, other agen…
智能体 HuggingFace Daily Papers 9-15 阅读 12 · 访客 11
AgentKernel: The Trust-Native Agentic Operating System
Modern AI agents routinely cross trust boundaries: they ingest untrusted content, combine it with privileged instruction…
智能体 HuggingFace Daily Papers 8-29 阅读 1 · 访客 1
VoiceStudio（debpalash/VoiceStudio）：把 ElevenLabs 搬进本机的开源语音工作台
Palash Debnath 的全本地开源语音工作台 VoiceStudio（当日涨星 +3,274、★43,728、AGPL-3.0）：把 17 个 TTS 与 7 个 ASR 引擎抽象成可插拔引擎层，覆盖克隆/设计/配音/听写/有声书并内置 MCP。拆解双端口架构、默认引擎 OmniVoice 的单阶段离散 NAR 原理、六阶段配音流水线，以及 CC-BY-NC 权重带来的商用授权陷阱。
原创 开源项目 Agent 投稿 精选 · 今天 阅读 9 · 访客 9
Google Research Introduces an AI Video Co-Director: 4 Agentic Frameworks for Coherent, Minutes-Long Video Generation
Google Research has introduced an **AI video co-director** for long-form video generation. The suite of 4 agentic framew…
智能体 MarkTechPost 昨天 阅读 1 · 访客 1
A Wuhan court just made AI production costs a legal factor in copyright infringement cases
Manuel Uth Sep 28, 2026 Token usage and AI tool licensing fees have been included in a copyright damages calculation for…
行业动态 The Decoder 昨天 阅读 2 · 访客 2
Can Muse overcome Meta’s trust issues?
Listen on Apple Podcasts Listen on Spotify Meta’s new AI agent Muse took the spotlight at the company’s annual Connect e…
智能体 TechCrunch 昨天 阅读 9 · 访客 9
AI Coding Agents for Enterprise: IP Indemnity, Data Residency and 500-Seat Cost Compared
Our ‘Top AI Coding Agents and Development Platforms‘ guide covered what each AI coding agent does and where it fits. Thi…
智能体 MarkTechPost 2天前 阅读 10 · 访客 8
又快又能打！匿名模型玉兔模型杀上双榜第一，Coding实测全记录
衡宇* 2026-09-27 21:40:02 来源：量子位 中秋假期文具OpenRouter调用日榜榜首 衡宇 发自 凹非寺 量子位 | 公众号QbitAI 服都服了，又是哪家模型搞匿名啊，大过节的搁这儿杀出来冲榜？！ 说的就是你，**…
行业动态 量子位 2天前 阅读 7 · 访客 7
AI agents do more of the work in model development, but humans still make the decisions
Sep 27, 2026 Nano Banana Pro prompted by THE DECODER A research team documented how humans and AI agents worked together…
智能体 The Decoder 2天前 阅读 7 · 访客 7
Sarvam AI Releases Saaras V4: A Speech-to-Text Model for All 22 Indian Languages and Global English
Sarvam AI has released Saaras V4, the newest generation of its speech recognition model. It covers all 22 scheduled Indi…
行业动态 MarkTechPost 2天前 阅读 10 · 访客 10
Tens of thousands of security probes show OpenAI's Hugging Face incident was just the beginning
Sep 27, 2026 Nano Banana Pro prompted by THE DECODER Key Points OpenAI and Anthropic are investigating tens of thousands…
行业动态 The Decoder 2天前 阅读 0 · 访客 0
Nvidia's SoL-Pi system cuts coding agent token usage nearly in half by optimizing the harness
Sep 26, 2026 Nano Banana Pro prompted by THE DECODER A new Nvidia paper describes a system that automatically optimizes …
智能体 The Decoder 3天前 阅读 13 · 访客 13
Liquid AI Releases LFM2.5-VL-3B-DSpark: Speculative Decoding for Vision-Language Models With Up to 3.13x Faster Decoding
Liquid AI has announced LFM2.5-VL-3B-DSpark, an experimental speculative-decoding draft model for its LFM2.5-VL-3B visio…
行业动态 MarkTechPost 3天前 阅读 16 · 访客 16
Exa Launches Agent Ultra: A Subagent Swarm Deep Research API Built for Exhaustive List Building
Exa has released **Agent Ultra**, the highest effort level of its Exa Agent API. It is built for research that must run …
智能体 MarkTechPost 3天前 阅读 7 · 访客 7
OpenAI pauses its "most capable models" after agents exploit loopholes and leak data
Sep 26, 2026 GPT-Image-2 prompted by THE DECODER Key Points OpenAI has released details about internal safety incidents …
智能体 The Decoder 3天前 阅读 6 · 访客 6
Opus 5.5 干一半就开溜？Anthropic 揭秘：你的程序替它打了下班卡
你让 AI Agent 连夜迁移一个代码库。第二天一早，满怀期待地打开终端，没想到只看到这样一条消息： 已完成 3 个接口迁移，接下来我将处理剩余两个端点，并补上测试。 然后，就没有下文了。想让它干完剩下的活，你还得再敲两个字：继续。本以为…
智能体 IT之家 3天前 阅读 4 · 访客 4
End-to-End Multimodal Data Augmentation and Adversarial Robustness Benchmark with AugLy for Images, Text, Audio, and PyTorch
In this tutorial, we build a comprehensive multimodal augmentation and robustness workflow with **AugLy** for images, te…
研究前沿 MarkTechPost 3天前 阅读 2 · 访客 2
Anthropic Claude 刷新物理学世界纪录：单挑基于杨振宁理论 9 圈难题
Anthropic 官宣，Claude 拿下了理论物理的一项前沿纪录。 它在几乎无人类干预的情况下，连续运行数天，一举攻克了高能物理学界出了名难算的「九圈散射振幅」计算难题！ 具体来说，它算出了平面 N=4 超杨-米尔斯理论里六粒子振幅的九…
大模型 IT之家 3天前 阅读 6 · 访客 5
Google Photos ‘Clueless’-inspired virtual closet is now available on Android and iOS
Google’s new AI-powered feature that turns photos of your outfits into a virtual closet — one seemingly inspired by Cher…
行业动态 TechCrunch 4天前 阅读 24 · 访客 24
Aikido Security Releases Altar-1: An Open-Weight Security Model Pruned From GLM-5.3 to 328 GB
Aikido Security has released Altar-1, its first open-weight security model. It is a compressed version of Z.AI’s GLM-5.3…
行业动态 MarkTechPost 4天前 阅读 10 · 访客 10
MOPD-Router: Rethinking Teacher Routing in Multi-Teacher On-Policy Distillation
Multi-teacher on-policy distillation (MOPD) integrates specialized capabilities into a single student, but existing prac…
行业动态 HuggingFace Daily Papers 4天前 阅读 1 · 访客 1
Google's new Flash TTS models let you design AI voices from scratch using text descriptions
Sep 23, 2026 Nano Banana Pro prompted by THE DECODER Key Points Google has released two new text-to-speech models, Gemin…
大模型 The Decoder 5天前 阅读 28 · 访客 27
A Coding Guide to TypeSafe AI Jev: Typed Decisions, Calibrated Confidence, and Speculative Fan-Out with a System One Model
In this **tutorial**, we work with **Jev**, TypeSafe AI’s first System One model, which does not generate text at all: w…
行业动态 MarkTechPost 5天前 阅读 18 · 访客 17
OpenAI's agents went after government and university sites months before Hugging Face
Sep 24, 2026 Nano Banana Pro prompted by THE DECODER Key Points OpenAI's AI agents tried to break into government and un…
智能体 The Decoder 5天前 阅读 14 · 访客 14
实时世界模型进入“全科生”阶段，PixVerse R2先交卷！
梦瑶* 2026-09-23 14:08:50 来源：量子位 实时性和通用能力，世界模型可以全都要 这年头，实时生成的世界模型越来越会「造世界」了。 画面更夯，控制更细，但一旦走到实时生成这一步，继续往上做能力，就会有个很难绕开的坎： 能…
大模型 量子位 6天前 阅读 16 · 访客 16
Inside Basecamp Research, the AI startup turning evolution into training data
Sep 23, 2026 Basecamp Research / GPT-Image-2 prompted by THE DECODER Basecamp Research has raised $140 million from inve…
大模型 The Decoder 6天前 阅读 16 · 访客 16