一篇读懂训练稳定性：Loss Spike、梯度爆炸与数值问题的诊断与修复
从 PaLM 与 OPT 的真实病历出发，拆解 loss spike 的五类根因、回滚跳批的标准操作与 z-loss、软截断等预防手段。
阅读全文 →
全部文章
ARCHIVE 共 465 条Don’t be fooled by this summer of AI hype
It’s been a busy few months for AI hype. At the end of April, Anthropic claimed that its model Claude Mythos is better a…
Alibaba Qwen Releases Qwen-Image-2.1: A 7B Open-Weight Model for Image Generation and Editing
Alibaba's Qwen team has released Qwen-Image-2.1, a 7B diffusion transformer that handles text-to-image generation, multi…
Xiaomi's affordable flagship AI leads the open models, and Anthropic says Claude helped get it there
Sep 22, 2026
Xiaomi
Key Points
Xiaomi's new MiMo-V2.6-Pro model leads current rankings of open AI models while costing f…
Flash-dLLM: IO-Aware KV Caching and Parallel Decoding for Fast, Memory-Efficient Diffusion LLMs
Diffusion Large Language Models (dLLMs) have recently emerged as a promising alternative to autoregressive LLMs by enabl…
后摩智能确认其下代大模型端边 AI 芯片采用 3D CIM 存算一体架构
IT之家 9 月 22 日消息，后摩智能 (HOUMO.AI) 在近日举行的 2026 全球 AI 芯片峰会上确认，该公司**下代大模型端边 AI 芯片将采用 3D CIM 存算一体架构**，结合融合存算与 3D DRAM 技术。
这款新一…
拿下 2 亿日活跃用户后，消息称字节跳动豆包收缩对话团队
IT之家 9 月 22 日消息，《晚点 LatePost》今天（22 日）晚间发布消息称，豆包通用 Session 团队正在减员，规模预计减半。该团队此前约有 50 人，**一度是豆包内规模最大的团队**，主要负责豆包的对话（Chat）产品…
Calibration as a First-Class Criterion in LLM Evaluation
Calibration of language models -- the alignment between expressed or implicit confidence and empirical correctness -- is…
JEV-as-a-Judge: Accept When Confident, Escalate When Unsure
LLM-as-a-judge enables evaluation across diverse tasks, but inference cost and confidence reliability become critical at…
A tiny software layer from lab-grown neurons promises faster, cheaper AI video
Sep 22, 2026
TBC / GPT-Image-2 prompted by THE DECODER
The Biological Computing Co. is teaming up with AWS to sell a tex…
Google confirms Gemini models hacked three companies in May 2026
A third-party cybersecurity firm accidentally gave experimental Gemini models access to the Internet.]
啊啊啊GPT-6 Astra这么不安全！这次马斯克都瘫坐了
97%尝试危险行为]
通过广告收集器 ChatGPT 知道你在其它网站的活动]
OpenAI 在 bzr.openai.com 上的广告收集器会设置名为 __obi、作用域为 .openai.com 的 Cookie。该 Cookie 的值在用户访问 ChatGPT 期间生效，并与用户的 ChatGPT 账户绑定。随后…