https://ai-tokens.cn/news/news-meta-muse-20260922.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Meta 个人 AI 智能体 Muse 上线两周登顶 App Store，单日市值飙升近 2000 亿美元 https://ai-tokens.cn/news/news-gpt-6-astra-20260903.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 OpenAI 发布 GPT-6 Astra：「欢迎来到 AGI 时代」，API 定价 $10/$50 为上代 2.5 倍 https://ai-tokens.cn/news/news-gemini-38-flash-20260903.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Google 三周内再推 Gemini 3.8 Flash：介绍价 $0.75/$3.75，Cyber 网络安全变体齐发 https://ai-tokens.cn/news/news-workbuddy-open-platform-20260903.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 WorkBuddy 开放平台解析：用 Skill、Connector、专家与知识库给 AI 同事「配工具」 https://ai-tokens.cn/news/news-wechatpay-ai-card-20260831.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 微信支付 AI 专属卡上新：一张卡直连 DeepSeek Harness 与 OpenClaw https://ai-tokens.cn/news/news-anthropic-mhs-20260828.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Anthropic 发布 MHS 模型硬件标准：让 AI 像插 USB-C 一样操控真实世界设备 https://ai-tokens.cn/news/news-dfd02e8c10946ab0.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 又一国产决策模型发布！准确率超过微软和Jev等决策模型 https://ai-tokens.cn/news/news-de46add1373038e9.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 既能“做选择题”，也能“做填空题”——星辰意图理解模型Xing4.0-29B-Intent发布 https://ai-tokens.cn/news/news-489b7bb699dc1c0f.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 AI 漫剧越做越快，赚钱却没那么容易了 https://ai-tokens.cn/news/news-fe786b5104d0ead4.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 IBM AI Releases Granite Speech 5.0 2B for Multilingual Transcription and Translation https://ai-tokens.cn/news/news-a7a75b8f97ef7762.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 How Far Can the Gemma 4 Speech Encoder Go Without Changing Its Weights? Saudi and Gulf Arabic Speech Recognition with a Frozen and an Adapter-Augmented Encoder https://ai-tokens.cn/news/news-3ef310cbc0f6a8c1.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 SkillEval: Learning Interpretable Ability Profiles of LLMs via Cognitive Diagnosis Models https://ai-tokens.cn/news/news-8dfa7ef84d604711.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Scaling Therapy Supervision: Assessing Psychotherapy Fidelity Using Large Language Models https://ai-tokens.cn/news/news-440cf8eaec9bd58c.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Is Anyone Enforcing the Cartel? What Reinforcement Learning Teaches LLM Agents in Markets https://ai-tokens.cn/news/news-1cf53f8f1a87fe63.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Where Do Legal Factors Live in LLMs? https://ai-tokens.cn/news/news-1597d96613c49911.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 LIFT: Scoped Constraints for Evaluating and Training Instruction Following https://ai-tokens.cn/news/news-96a3ee7df5835eea.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 How to Use Parallel Data for Multilingual Pretraining https://ai-tokens.cn/news/news-2bb631880a190817.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 SANE: Automated Scorer-Specific Metamorphic Testing for LLM Judges https://ai-tokens.cn/news/news-66f56984b6446ca0.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Building AI Research Teams for Autonomous Discovery of Linear Attention Architectures https://ai-tokens.cn/news/news-0fe2424ead31ec85.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 CSD: Common-Support Distillation for Cross-Tokenizer Knowledge Transfer https://ai-tokens.cn/news/news-91a45a8a5c65eebd.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 JaxMARL-BC: Multi-Agent Reinforcement Learning for Macroeconomic Modelling, in Seconds https://ai-tokens.cn/news/news-3e257f94c6eaa097.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 The Butterfly Effect in LLM Social Simulations: From Architecture Sensitivity to Robustness Audits https://ai-tokens.cn/news/news-7cff9a65cdd4fdbf.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Parallel AI Agent Collaboration as Graph-Structured Distributed Block Optimization https://ai-tokens.cn/news/news-9050b7c4e37f6fb7.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 To Manipulate or Not: Opinion Dynamics in LLM Interactions https://ai-tokens.cn/news/news-7d39b13be0bc98a9.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Transformers as Generic Deep Neural Networks: Exact Compilation, the Cost of Removing Activation-Activation Products, and Multi-Agent Attention Clustering https://ai-tokens.cn/news/news-68b1885591648124.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Resonance and Depth: Composing Timbral Identity and Spatial Form in an Artist-in-the-Loop Co-Creative Multi-Agent Percussive Performance https://ai-tokens.cn/news/news-5b6476003a786997.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 MA-UITD: Offline Multi-Agent Reward Inference for Characterizing Continuous Social Behavior https://ai-tokens.cn/news/news-7f07b7a2a2a2c7ff.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Design and Tuning of Multi-Agent Systems Using Bayesian Evolutionary Optimization https://ai-tokens.cn/news/news-c64e9f961cf3b2c8.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Credit Where Credit Is Due: Contrastive Attribution Topology Search for Multi-Agent Systems https://ai-tokens.cn/news/news-61c0403eaea2594c.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 OACap: Caption-Guided Active Observation for Detailed Video Captioning https://ai-tokens.cn/news/news-63d0fbb89ed9cb8c.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Conflicted Records: Typed Adversarial Contracts for Auditable LLM Output Quality Enforcement https://ai-tokens.cn/news/news-7af4030312fdc9aa.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 ToM2Vec: Continuous Theory-of-Mind Tokens for Multi-Agent Coordination https://ai-tokens.cn/news/news-d8e1fb9d8f652d16.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 To Keep or Not to Keep: Learning KV Cache Retention in Disaggregated LLM Serving Systems https://ai-tokens.cn/news/news-fe44ca1df228045d.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Reinforcement learning for hardware acceleration and system-level optimization on FPGA and integrated-circuit platforms: A survey https://ai-tokens.cn/news/news-abf61122a19388ab.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Scalable Stress-Constrained Topology Optimization: A Super-resolution and Simultaneous Approach Based on Machine Learning https://ai-tokens.cn/news/news-f6cdbe2f16bc3202.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 SimHack: Predicting the onset of detectable reward hacking before RL training https://ai-tokens.cn/news/news-e513ec69bf8cf41b.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 STAR-XPC: Set Transformer for Estimating X-ray Transmission in Polymer Composites https://ai-tokens.cn/news/news-28e14cad8781cc9e.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Offline Apprenticeship Learning for Human-Centric Environments https://ai-tokens.cn/news/news-e39d96f2d8aafc8c.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 ReAPO: Retry-Aware Policy Optimization for Efficient Exploration in LLM Reasoning https://ai-tokens.cn/news/news-c66559a1ec71f858.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Building a Nested LADDER for LLM Unstructured Post-Training Pruning https://ai-tokens.cn/news/news-a9a53b5160aee3d4.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Second Thoughts: Exploring and Planning through Comparative Imagination in World Models https://ai-tokens.cn/news/news-d12b52d75eabb204.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Extreme Low-State Quantized Transformer for Electricity-Load Forecasting https://ai-tokens.cn/news/news-03aea1bc87026e4b.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Measure First, Then Allocate: Calibrated Rollout Budgets for MCP Agents in Reinforcement Learning https://ai-tokens.cn/news/news-be9fe8b1a196562d.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 ArcFlow: Training Flow Policies for Multiple Solver Budgets https://ai-tokens.cn/news/news-266ab0d2c3042c9f.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 OpenAI agents analyze mouse brain data in 32 minutes, a job that would take humans months https://ai-tokens.cn/news/news-1b5147a9ed248a83.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Anthropic takes on Grok Bot, Meta's Muse, and OpenAI's Dots with its always-on agent "Claude Duet" https://ai-tokens.cn/news/news-5ba1ec3db181cdc2.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 AI’s hidden storage bill: What inference leaves behind https://ai-tokens.cn/news/news-46751b8492c33b4b.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 全球算力价格指数启动 九章云极核心共建算力定价基准 https://ai-tokens.cn/news/news-69994ae66cb875d1.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 10-50万级全覆盖，千里智驾量产进入放量期 https://ai-tokens.cn/news/news-339b8d911d2d7a0f.html AI导航 zh-CN 2026-10-22T19:00:03+08:00 Perplexity and NVIDIA Researchers Introduce JOLT: a Joint Teacher-Student Reinforcement Learning Method