Xiaohu AI Daily
Every day 10:00 · Curated AI Updates
中 EN 日
2026-05-30 · Sat generated 10:26:34
Sources
184
Items
800
Score 8+
19
Clusters
4
🌟 Today's Headline
Anthropic rolls out Claude Opus 4.8 with near-Mythos level alignment and 3x cheaper fast mode
Anthropic launched Claude Opus 4.8, a new flagship model that prioritizes reliability over raw performance. The model introduces a five-tier Thinking effort selector allowing users to balance computation and output quality. Opus 4.8 scores 88.6% on SWE-bench Verified and 74.6% on Terminal-Bench 2.1, outperforming GPT-5.5 and Gemini 3.1 Pro. Its defining feature is reduced likelihood of silently approving flawed code—four times lower than version 4.7—while actively flagging uncertainties and questioning unsupported assumptions. A new Fast Mode delivers 2.5x faster output with significantly lower API pricing: $10 per million input tokens and $50 per million output tokens. Dynamic Workflows in Claude Code enable single prompts to spawn multi-agent teams for complex tasks. This release signals a strategic shift in frontier model development from capability maximization to trust and alignment.
💬 Editor's Note
Anthropic is playing a different game: betting on reliability over benchmark scores. Opus 4.8's refusal to silently pass bad code and admission of uncertainty matter more to real-world developers than marginal performance gains.
Read more → Product
🔥Today's Highlights
01
Anthropic Raises $65B at $965B Valuation, Becomes World's Most Valuable Startup
10/10
Anthropic has completed a historic $65 billion funding round at a $965 billion valuation, making it the world's most valuable startup and officially surpassing OpenAI in market value. The funding was led by Greenoaks, Sequoia, Altimeter, and Dragoneer, with strategic new investors including semiconductor giants Samsung, Micron, and SK Hynix joining the round.
Read more →
02
9 demos of Gemini Omni and Gemini 3.5 in action
9/10 New Product
Google showcases Gemini Omni and Gemini 3.5 through nine live demonstrations highlighting multimodal capabilities, including real-time video understanding, speech interaction, and cross-modal reasoning. The demos illustrate the models' practical applications across different use cases.
Read more →
03
OpenAI gives GPT-5.5 Instant a readability upgrade while phasing out two older models
9/10 New Product
OpenAI updates GPT-5.5 Instant for more natural, human-like responses, discontinues Canvas feature by moving writing and coding tasks directly into chat. The company also retires older o3 and GPT-4.5 models from ChatGPT, streamlining the available model lineup.
Read more →
04
v0.30.0
9/10 New Product
Ollama v0.30.0 restructures the underlying architecture to directly support llama.cpp instead of GGML, enabling full GGUF file format compatibility. MLX acceleration on Apple Silicon is integrated to improve inference performance on Mac devices.
Read more →
05
After Nvidia’s $20B not-acqui-hire, AI chip startup Groq reportedly raising $650M
9/10 News
Chipmaker Groq is looking to raise $650 million in internal funding as it pivots from hardware to focus more on AI inference, the process of refining the way AI models respond to prompted requests, per Axios.
Read more →
06
TeamCity 2026.1.1 Is Now Available
9/10 News
Today we’re rolling out the first bug-fix for TeamCity On-Premises 2026.1 servers. This update addresses over 20 issues and performance issues, including: See TeamCity 2026.1.1 Release Notes for the complete list of resolved issues. Why update? Staying up to date with minor releases ensures yo
Read more →
📊Topic Clusters
📌 大模型版本发布周期
Anthropic、Google、OpenAI 等主流厂商的新模型版本发布与性能迭代
Anthropic rolls out Claude Opus 4.8 with near-Mythos level alignment and 3x cheaper fast mode 10
9 demos of Gemini Omni and Gemini 3.5 in action 9
OpenAI gives GPT-5.5 Instant a readability upgrade while phasing out two older models 9
v0.30.0 9
Qwen-VLA: Unifying Vision-Language-Action Modeling across Tasks, Environments, and Robot Embodiments 6
📌 AI 融资与企业商业化
头部 AI 公司融资轮次、估值提升与商业化订阅计划推出
Anthropic Raises $65B at $965B Valuation, Becomes World's Most Valuable Startup 10
After Nvidia’s $20B not-acqui-hire, AI chip startup Groq reportedly raising $650M 9
Meta Launches Tiered Subscription Model Across Apps with Paid AI Features 8
📌 AI 产品功能与开发者工具
各平台新增功能（对话管理、多语言支持、文件生成等）与开发者工具升级
Now you can use your OpenRouter models directly inside @ComfyUI workflows！ 9
codex for managing the codex UI： 9
For every ChatGPT conversation that started as "one quick thing" and became a full on saga： table of… 9
The blog post did the thinking. Now let the promo do the work. Drop in the content. Define the hoo… 9
We are continually adding new models and endpoints to the Runway API so you can build the best gener… 9
OpenAI for realtime translation - speak in any of 70+ input languages and translate into 13 output o… 9
Windows users， this one's for you. Computer use now works on Windows， so Codex can take action on y… 9
OpenRouter now supports "apply_patch，" a server tool that lets any model propose file edits using V4… 9
v2.1.157 6
📌 API 服务质量与计费透明度
模型服务的配额管理、超额扣费问题、成本控制的行业困局
One company reportedly spent $500 million on Claude in one month after failing to cap AI usage 7
Token Inflation: How Dishonest Providers Can Overcharge for Large Language Model Usage 6
Google fixes several bugs in Gemini usage limits that burned through quotas too fast 6
📖Worth a Deep Read
🕐 ~9 min read · Industry 8/10
Meta Launches Tiered Subscription Model Across Apps with Paid AI Features
💡 Industry trends and analysis
Meta has officially launched tiered subscription services across Instagram, Facebook, WhatsApp, and Meta AI under a unified "Meta One" brand, marking a significant shift in the company's core business model. The subscription offerings include Instagram Plus and Facebook Plus at $3.99/month with customization features and enhanced analytics; WhatsApp Plus at $2.99/month for advanced functionality; and two Meta AI tiers—Meta One Plus ($7.99/month) and Premium ($19.99/month), with Premium tier offering faster "thinking mode" responses for complex queries. Additional creator and business subscription options are being tested with verification badges, expanded promotional tools, and analytics capabilities. This strategic diversification reflects Meta's escalating AI infrastructure costs, with the company committing up to $145 billion to AI in 2026 alone, requiring new revenue streams beyond advertising to fund massive AI investments.
Read more →
🕐 ~3 min read · Tutorial 7/10
Take our I/O 2026 quiz， vibe coded in Google AI Studio.
💡 Can be adapted into tutorial material
Google 使用其开发工具 Google AI Studio，通过氛围编程（vibe coding）方式，创建了一个关于 Google I/O 2026 主要公告的在线测验。
Read more →
🕐 ~3 min read · Tutorial 7/10
What happens when companies become too AI-pilled？
💡 Can be adapted into tutorial material
Box创始人Aaron Levie指出，决定用AI替代员工的人往往最不了解工作的实际内容，他将此称为"AI psychosis"。ClickUp近期为部署AI智能体裁员22%即是一例。2026年的科技行业裁员规模已接近2025年全年。
Read more →
🕐 ~3 min read · Tutorial 7/10
这个 skill 看着不错，可将文字、URL 或文章直接生成公众号首图、小红书图文卡、教程步骤卡等视觉物料，支持 28 种布局和 10 种主题。
💡 Can be adapted into tutorial material
claude-design-card 是一款专为中文内容创作者设计的 Skill。它能将文字、URL 或文章直接转化为可发布的视觉卡片，如公众号首图、小红书图文卡、教程步骤卡等，支持 28 种布局与 10 种主题。其核心价值在于自动化了"写完文章"后最繁琐的流程：自动提炼重点、选择版式、生成 HTML 并截图成 PNG，替代了以往手动使用 Figma 或 Canva 等工具的步骤。该工具开源，适合经常撰写相关内容的创作者尝试。
Read more →
🕐 ~3 min read · Tutorial 7/10
The team at @llama_index built an awesome template using LlamaParse and the new Managed Agents in th…
💡 Can be adapted into tutorial material
LlamaIndex 团队基于 Google 新发布的 Agents API 构建了一个模板，使智能体能够访问 LlamaParse 和 LiteParse，从而自动处理非结构化文档。其工作流程为：配置数据与输出的 Git 仓库，将仓库克隆至智能体沙箱，安装 LiteParse CLI 与 LlamaParse SDK 及相关技能，最后通过提示词驱动智能体自主执行任务。该模板最终形成一个可直接使用 LlamaParse 和 LiteParse 处理真实世界文档的智能体。
Read more →
📂Browse by Category
New Product
Now you can use your OpenRouter models directly inside @ComfyUI workflows！
9
现在你可以直接在ComfyUI工作流中使用你的OpenRouter模型了！ 【引用 @ComfyUI】：ComfyUI刚刚添加了@OpenRouter支持。 你不再局限于单一的大语言模型，现在可以直接在Comfy中访问20多个模型。 更多灵活性，更少摩擦，同样的工作流。 工作流链接在下方👇
Read more →
codex for managing the codex UI：
9
Codex用于管理Codex界面： 【引用 @guinnesschen】：如果你厌倦了管理Codex对话线程，就让Codex自己管理自己吧！Codex现在可以创建对话线程、搜索它们、整理它们、固定重要的线程，并为并行任务启动工作树。
Read more →
For every ChatGPT conversation that started as "one quick thing" and became a full on saga： table of…
9
对于每个始于"就问一件事"却演变成完整长篇的ChatGPT对话：目录功能现已推出。 适用于包含5条以上回复的对话。
Read more →
Opinion
Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet
6
Anthropic researchers demonstrate that sparse autoencoders can extract interpretable features from Claude 3 Sonnet at production scale, with up to 34 million features extracted from the model's residual stream. The breakthrough shows dictionary learning methods scale beyond small transformers.
Read more →
First head-to-head comparison of agentic AI applied to the analysis of simulated data of the Einstein Telescope
6
Researchers conduct the first head-to-head benchmark comparing Claude Code (Anthropic) and Codex (OpenAI) on autonomous gravitational wave data analysis pipelines. Both agentic systems execute tasks without human intervention on shared infrastructure, revealing performance differences in complex scientific workflows.
Read more →
Evolutionary Dynamics of Cooperation in Next-Generation LLM Agent Systems: A Cross-Provider Empirical Extension
6
Extension of Willis et al.'s evolutionary game theory benchmark (Iterated Prisoner's Dilemma) to newer frontier models, investigating whether larger, diverse LLMs retain the cooperative biases observed in ChatGPT-4o and Claude 3.5 Sonnet or exhibit different equilibrium behavior.
Read more →
Industry
UEFA and UC3 have announced Alibaba Group as global AI， Cloud Computing and E-Commerce Partner for U…
7
阿里云和Qwen成为UEFA官方独家AI、云计算与电子商务合作伙伴，合作期覆盖2027/2028赛季至2032/2033赛季的UEFA男子俱乐部赛事，以及UEFA EURO 2028。阿里巴巴集团主席蔡崇信表示，将投入云计算、全栈AI及全球电商平台能力，支持赛事运营。
Read more →
Trends in AI and Human-AI Interaction in Clinical Trials -- A Hybrid Human-AI Exploration
6
Study analyzes ClinicalTrials.gov records to track temporal trends in AI terminology usage and geographical distribution of AI-driven clinical trials. Researchers employed GPT-5.5 combined with human review to systematically characterize human-AI interaction patterns in medical research.
Read more →
FormInv: A Measurement Protocol for Semantic Invariance in Mathematical Reasoning Benchmarks
6
Audit of MathCheck benchmark identifies 4 semantically flawed paraphrases (3.1% of test set), causing ranking volatility—GPT-4o drops from rank 2 to 4, while Claude Haiku and DeepSeek V3 rise above it. Cross-model consensus (≥3/4 models) detected errors automatically at minimal cost.
Read more →
Tech
I'm very excited by this new benchmark dataset for visual generation that is suitable for the modern…
7
我对这个适用于大规模生成模型新时代的视觉生成基准数据集感到非常兴奋！🤩
Read more →
Tutorial
How Braintrust turns customer requests into code with Codex
9
Braintrust engineers leverage Codex and GPT-5.5 to accelerate code generation for customer-facing features. The case study demonstrates how AI-assisted coding reduces development time and enables faster experimentation cycles for building new capabilities.
Read more →
Cognition's Scott Wu says AI coding agents shouldn't replace humans
7
Cognition公司开发了Devvin，这是一个号称首个且最成功的AI编程智能体。其著名程序员创始人Scott Wu明确表示，该智能体并非旨在取代人类程序员。
Read more →
I had to test it myself to believe this unreal inference speed. 3，000 tokens/s for 1 user on standa…
7
Kog团队在标准数据中心GPU上实现了极高的单用户推理速度，在8× AMD MI300X GPUs上达到3，000 tokens/s，在8× NVIDIA H200上达到2，100 tokens/s。相比常规推理速度（约100-300 tokens/s），实现了10-30倍提升。
Read more →
📭Skip Today

Auto-filtered. Here's why — so you know you're not missing out:

📎 Long Tail (239) · click to expand
Behavior-Induced Mirror-Prox Temporal-Difference Learning for Faster Off-Policy Prediction 5
MemoSight: Unifying Context Compression and Multi Token Prediction for Reasoning Acceleration 5
Hierarchical Task Network Planning with LLM-Generated Heuristics 5
Databricks at SIGMOD 2026 5
[AINews] Founders and Forward Deployed Engineers 5
Frontier LLM-based agents can overcome the ontology curation bottleneck for natural phenotypes 5
VFEAgent: A Multimodal Agent Framework for End-to-End Automated Finite Element Analysis 5
BEAMS: Benchmarking and Evaluating AI for Modeling and Simulation 5
Practitioner Beliefs and Behaviors in AI-Enhanced Education: DOT Framework Survey Evidence 5
Better Later Than Sooner: Neuro-Symbolic Knowledge Graph Construction via Ontology-grounded Post-extraction Correction 5
Tailoring the Curriculum: Student-Centered Reasoning Distillation via Dynamic Data-Model Compatibility 5
Surfacing Isolated Learners with Outcome-Independent Mediation of Feedback between Teachers and Students Using AI 5
Diagnosing Harmful Continuation in Answer-Correct Long-CoT Training Traces 5
EvoMD-LLM: Learning the Language of Species Evolution in Reactive Molecular Dynamics 5
Battery-Sim-Agent: Leveraging LLM-Agent for Inverse Battery Parameter Estimation 5
Planning with the Views via Scene Self-Exploration 5
GPS-Enhanced Tourist Mobility Modeling with Seasonal Spatial Priors and LLM-Based Activity Chain Generation 5
BitTP: The Lightweight Trajectory Prediction Model with BitLLM for Edge-Devices 5
Uncertainty-Aware Transfer Learning for Cross-Building Energy Forecasting: Toward Robust and Scalable District-Level Energy Management 5
Benchmarking Positional Encoding Strategies for Transformer-Based EEG Foundation Models 5
From XXLTraffic to EvoXXLTraffic: Scaling Traffic Forecasting to Sensor-Evolving Networks 5
Croissant Tasks: A Metadata Format for Reproducible Machine Learning Evaluations 5
PRAIB: Peer Review AI Benchmark of Behaviour of LLM-Assisted Reviewing 5
OptSkills: Learning Generalizable Optimization Skills from Problem Archetypes via Cluster-Based Distillation 5
Compass: Navigating Global Marine Lead Data Integration through Expert-Guided LLM Agent 5
Conformal Certification of Reasoning Trace Prefixes 5
Temporal Stability and Few-Shot Prompting in Math Task Assessment 5
S3Mem: Structured Spatiotemporal Scene-Event Memory for Long-Horizon Interactive Question Answering 5
A comparative study of transformer-based embeddings for topic coherence 5
The Hamilton-Jacobi Theory of Deep Learning 5
Specialty-Specific Medical Language Model for Immune-Mediated Diseases 5
Comparing Post-Hoc Explainable AI Methods for Interpreting Black-Box EEG Models in Depression Detection 5
LoRe: Adaptive Interaction-Evaluation Routing with Per-Step Interaction Budgets for Iterative Graph Solvers 5
Stochastic Lifting for Generating Trajectories of Stochastic Physical Systems 5
Toward Ethical Facial Age Estimation: A Generalized Zero-Shot Benchmark Without Training on Children's Data 5
Wait! There's a Way Out: A Decision Mechanism for Forecasting Conversational Derailment 5
KLAS: Using Similarity to Stitch Neural Networks for Improved Accuracy-Efficiency Tradeoffs 5
Causal Label Recovery in Payment Networks 5
Pocket-Dentist: On-Device Dental Image Understanding via Efficient Multimodal Large Language Models 5
Rethinking FID Through the Geometry of the Reference Dataset 5
The Good, the Bad, and the Ugly of Markov Boundary for Tabular Prediction 5
Evolutionary Rule Extraction from Corporate Default Prediction Models 5
Network Optimization Aspects of Autonomous Vehicles: Challenges and Future Directions 5
Temporal Motif-aware Graph Test-time Adaptation for OOD Blockchain Anomaly Detection 5
Singularity-aware Optimization via Randomized Geometric Probing: Towards Stable Non-smooth Optimization 5
Brain-IT-VQA: From Brain Signals to Answers 5
Learning Context-Conditioned Predicate Semantics via Prototype Feedback 5
COMET: Concept Space Dissection of the Modality Gap in Audio-Text Multimodal Contrastive Embeddings 5
Predicting Causal Effects from Natural Language Queries using Structured Representations 5
Personalized Turn-Level User Conversation Satisfaction Benchmark 5
A unified deeplearning framework for contrast-phase-specific virtual monochromatic imaging 5
Energy-Aware NECO for Single-Pass Pixel-wise Out-of-Distribution Detection in Semantic Segmentation 5
Mitigating Stethoscope-Induced Shortcuts in Respiratory Sound Classification under Federated Domain Generalization with Causality-Inspired Interventions 5
Evaluating Skill and Stability of ArchesWeather and ArchesWeatherGen under Multi-Decadal Climate Simulations 5
Genetically Aligned Patient Representations Improve Hematological Diagnosis 5
xModel-KD: Cross-modal Knowledge Distillation for 3D Scene Perception using LiDAR 5
Evolving Features vs Evolving Entire Trees with GP for Interpretable Survival Analysis 5
Beyond MSE: Improving Precipitation Nowcasting with Multi-Quantile Regression 5
DAMEL: Dual-Axis Multi-Expert Learning for Class-Imbalanced Learning 5
On Distributional Reinforcement Learning in Chaotic Dynamical Systems 5
iLoRA: Bayesian Low-Rank Adaptation with Latent Interaction Graphs for Microbiome Diagnosis 5
What drives performance in molecular MPNNs? An operator-level factorial benchmark 5
HPO: Hysteretic Policy Optimization for Stable and Efficient Training under Sparse-Reward Regime 5
Before the Shutter: Aesthetic and Actionable Portrait Photography Planning in 3D Scenes 5
Improved Guarantees for Heterogeneous Treatment-Effect Estimation via Matrix Completion 5
In-Context Reward Adaptation for Robust Preference Modeling 5
On Language Generation in the Limit with Bounded Memory 5
PuzzleClone: A DSL-Powered Framework for Synthesizing Verifiable Data 5
EAPO: Enhancing Policy Optimization with On-Demand Expert Assistance 5
TANDEM: Temporal-Aware Neural Detection for Multimodal Hate Speech 5
Benchmarking at the Edge of Comprehension 5
Recurrent Structural Policy Gradient for Partially Observable Mean Field Games 5
Weakly Supervised Detection and Temporal Localization of Whale Calls in Long-Duration Bioacoustic Data 5
Page image classification for content-specific data processing 5
Scalable RF Simulation in Generative 4D Worlds 5
GRPO is Secretly a Process Reward Model 5
Towards Foundation Models for Zero-Shot Time Series Anomaly Detection: Leveraging Synthetic Data and Relative Context Discrepancy 5
The Impact of Semantic Pairs on Self-Supervised Representation Learning 5
Offline Reinforcement Learning with Generative Trajectory Policies 5
ScheduleStream: Temporal Planning with Samplers for GPU-Accelerated Multi-Arm Task and Motion Planning & Scheduling 5
An accuracy-aware extension to LRP-based pruning for CNNs to prevent cascading accuracy degradation in data-scarce transfer learning 5
The Best of the Two Worlds: Harmonizing Semantic and Hash IDs for Sequential Recommendation 5
A Review of Learning-Based Motion Planning: Toward a Data-Driven Optimal Control Approach 5
Learn from A Rationalist: Distilling Intermediate Interpretable Rationales 5
Beyond Normalization: Rethinking the Partition Function as a Difficulty Scheduler for RLVR 5
Rooted Absorbed Prefix Trajectory Balance with Submodular Replay for GFlowNet Training 5
AG-REPA: Causal Layer Selection for Representation Alignment in Audio Flow Matching 5
AuthorMix: Modular Authorship Style Transfer via Layer-wise Adapter Mixing 5
ReSpinQuant: Efficient Layer-Wise LLM Quantization via Subspace Residual Rotation Approximation 5
Causal Disentanglement-Inspired Degradation Representation Learning for Full-Reference Image Quality Assessment 5
Self-Supervised Laplace Approximation for Bayesian Uncertainty Quantification 5
Coarse-to-Fine Domain Incremental Learning with Attentive Distillation for Mining Footprint Segmentation in Multispectral Imagery 5
Theoretical Analysis of Sparse Optimization with Reparameterization, Weight Decay, and Adaptive Learning Rate 5
Autoregression-Free Neural Operators for Time-Dependent PDEs 5
Hands-On With Gemini Spark: I Gave It Access to My Life and It Friend-Zoned My Boyfriend 5
The Vatican’s Man Inside Anthropic 5
Kiwibit’s AI-powered bird feeder is my new backyard buddy 5
Comprehensive observability for Amazon SageMaker AI LLM inference: From GPU utilization to LLM quality 5
Cool stuff Google Cloud customers built, May edition: Agentic algorithms for supply chains; virtual try-on APIs; robotic camera operators & more 5
From petabytes to predictions: Easy BigQuery insights in Google Sheets 5
AI-Assisted Migration Tool Helps Teams Move from ingress-nginx to Higress in Minutes 5
Adobe’s conversational AI agent is a mediocre design intern 5
What Does It Actually Take for an IDE to Understand Rust? 5
Premium: What If...We're In An AI Bubble? (Part 3) 5
Cloud CISO Perspectives: How to build an AI-ready security program for the public sector 5
Soro: A Lightweight Foundation Model and Chatbot for Tajik 5
You Are in Control of Your State: Why Human Outcomes Are Controllable Through Causal State Intervention 5
Dataset-Driven Channel Masks in Transformers for Multivariate Time Series 5
NCSAM Noise-Compensated Sharpness-Aware Minimization for Noisy Label Learning 5
GICDM: Mitigating Hubness for Reliable Distance-Based Generative Model Evaluation 5
Relational In-Context Learning via Synthetic Pre-training with Structural Prior 5
ProtoMedAgent: Multimodal Clinical Interpretability via Privacy-Aware Agentic Workflows 5
Bridging Classification and Reconstruction: Cooperative Time Series Anomaly Detection 5
Prospective evaluation of multimodal respiratory failure prediction: Do chest X-rays improve performance beyond EHR signals? 5
The Alignment Floor: How Persona Customization Breaks Safety in Weakly-Aligned LLMs 5
BioELX: Cross-lingual Biomedical Entity Linking via Alias-based Retrieval and LLM Ranking 5
Simorgh at SemEval-2026 task 7: Region-Aware Hybrid Retrieval for Low-Resource Cultural Reasoning in Multilingual Question Answering 5
UNIQUE: Universal Top-k Sparse Attention for Training-free Inference and Sparsity-aware Training 5
Syllabic-Structure Decoder for Automatic Speech Recognition in Vietnamese 5
DisasterBench: Benchmarking LLM Planning under Typed Tool Interface Constraints 5
KVoiceBench, KOpenAudioBench, and KMMAU: Agent-Driven Korean Speech Benchmarks for Evaluating SpeechLMs 5
Knowledge Dependency Estimation for Reliable Question Answering 5
Challenges in Explaining Pretrained Clinical Text Classifiers 5
Chinese Word Boundary Recovery through Character Alignment Projection 5
Better heads do not guarantee better binarized constituency parsing 5
DEPART: DEcomposing PARiTy across Multilingual LLMs 5
BenGER: Benchmarking LLM Systems on Subsumption-Based Legal Reasoning in German Law 5
The Harder Text Embedding Benchmark (HTEB): Beyond One-dimensional Static Robustness 5
Analyzing Quality-Latency-Resource Trade-offs in a Technical Documentation RAG Assistant Using LoRA Adaptation 5
Why We Need Speech to Evaluate Speech Translation 5
When Seekers Are Hard to Help: Evaluating Emotional Support Dialogue Systems in Worst-Case Interactions 5
Revisiting Anthropomorphic Reflection Markers in Large Language Model Reasoning 5
HELEA: Hard-Negative Benchmark and LLM-based Reranking for Robust Entity Alignment 5
When Discourse Pressures Conflict: Information Structure in Vision-Language Model Outputs 5
PubMedCausal: A Span-Level Annotated Corpus for Causal Relation Extraction in Biomedical Text 5
A new semantically annotated corpus with syntactic-semantic and cross-lingual senses 5
On Compositional Learning Behaviours in Formal Mathematics 5
The Attentional White Bear Effect in Transformer Language Models 5
Sense Representations Are Inducible Interfaces 5
Stance Detection in Prediction Markets: Addressing Imbalanced Trader Commentary via Counterfactual Augmentation and Market Context 5
Rethinking Memory as Continuously Evolving Connectivity 5
The Abstraction Gap in Vision-Language Causal Reasoning 5
VLMs May Not Globally Enhance Human Alignment over LLMs During Natural Reading 5
Identifying and Understanding Human Values in Text: A Tailorable LLM-based Architecture 5
Memory-Based vs. Context-Only Conditioning Produces Distinct Behavioral Patterns in Stateful Personalization 5
REC-CBM: Rubric-Aware Error-Correction Concept Bottleneck Models for Trustworthy Open-Ended Grading 5
Generic Interpretation Approach for Transformer Models Incorporating Heterogenous Attention Structures 5
CAREF: Calibration-Aware Regularization for Explanation Faithfulness Without Rationale Supervision 5
VCap: Hypergeometric Rewards for Weak-to-Strong Visual Captioning 5
MIRA: A Bilingual Benchmark for Medical Information Response Audit 5
A Wolf in Sheep's Clothing: Targeted Routing Hijacking in Federated RAG 5
MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated Content 5
Explaining is Harder Than Predicting Alone: Evaluating Concept-based Explanations of MLLMs as ICL Visual Classifiers 5
Risk-Controlled Lean-as-Judge for Natural-Language Mathematical Reasoning 5
Satisfiability Solving with LLMs: A Matched-Pair Evaluation of Reasoning Capability 5
GraphSteal: Structural Knowledge Stealing from Graph RAG via Traversal Reconstruction 5
Activation Steering for Synthetic Data Generation: The Role of Diversity in Downstream Safety Detection 5
Extrapolative Weight Averaging Reveals Correctness-Efficiency Frontiers in Code RL 5
Personal Visual Memory from Explicit and Implicit Evidence 5
On the Fallacy of Global Token Perplexity in Spoken Language Model Evaluation 5
Attention Projection Mixing with Exogenous Anchors 5
PEAR: Pairwise Evaluation for Automatic Relative Scoring in Machine Translation 5
RMPL: Relation-aware Multi-task Progressive Learning with Stage-wise Training for Multimedia Event Extraction 5
Explanation Generation for Contradiction Reconciliation with LLMs 5
A tree interpretation of arc standard dependency derivation 5
Why Gaussian Diffusion Models Fail on Discrete Data and How to Prevent It? 5
Speaking of Language: Reflections on Metalanguage Research in NLP 5
Adaptive Cost-Efficient Evaluation for Reliable Patent Claim Generation 5
Compositional Consistency-Guided Decoding for Three-Way Logical Question Answering 5
Early Decisions Matter: Proximity Bias and Initial Trajectory Shaping in Non-Autoregressive Diffusion Language Models 5
Evaluating the Evaluator: Problems with SemEval-2020 Task 1 for Lexical Semantic Change Detection 5
BenGER Platform: A Collaborative Web Platform for End-to-End Benchmarking of German Legal Tasks 5
A Benchmark Construction and Evaluation Framework for Specialist Domains: Case Study on Defense-related Documents 5
Syntax as a Rosetta Stone: Universal Dependencies for In-Context Coptic Translation 5
Heterogeneous Dependency Graph-Guided Attentionfor Patent Representation Learning 5
FEA-SLT: A Gloss-Free End-to-End Framework for Facial-Expression-Aware Sign Language Translation 5
An Effective-Rank Audit of Alignment-Induced Activation Shifts: Confound Control, Constructive Calibration, and Limits 5
MerLean-Prover: A Recursive Looping Harness for Lean 4 Theorem Proving 5
Meta-Programming for Linear-time Temporal Answer Set Programming 4
Online Fair Division with Additional Information 4
Architecture-Induced Recoverability Bias in Differentiable Symbolic Regression 4
Winning under CMS TEAM: Building the learning health system to realize success in VBC today and tomorrow 4
Differentiable Belief-based Opponent Shaping 4
The Confidence Shortcut: A Reasoning Failure Mode of Masked Diffusion Models 4
Certified Policy Optimisation for Nested Causal Bandits via PAC-Bayes Risk 4
Quantifying and Optimizing Simplicity via Polynomial Representations 4
On the Geometry of Games and their Solvers 4
Self-Play Reinforcement Learning under Imperfect Information in Big 2 4
TaxDistill: Improving Metagenomic Taxonomic Annotation via Distilled Genomic Foundation Models 4
Domain-Informed Representation for Evolutionary Sieving in Integral and Module Lattices 4
Extreme dynamic symmetry enables omnidirectional and multifunctional robots 4
The Sample Complexity of Multiclass and Sparse Contextual Bandits 4
Selection Hyper-heuristics Can Automatically Adjust the Learning Period to Optimally Solve Pseudo-Boolean Problems 4
Obfuscation Rules for Detecting and Detoxifying Korean Toxicity 4
Topological Order in Neural Wavefunctions 4
AlloyDB Hot Standby: Faster failovers, consistent performance 4
How the Pope’s Magnifica Humanitas offers a template for individuals to meet the AI moment 4
We Asked the ‘Future of Truth’ Author to Explain How He Used AI. It Didn’t Go Well 4
The Tradeoff That Slows Production Teams Down: Flexibility vs Actually Shipping 4
How Wearable IoT Enables Real-Time Fall Detection and Alerts 4
Build Professional Web Scrapers That Actually Work 4
The find out stage of AI is just supply chain and password protection​​​​‌‍​‍​‍‌‍‌​‍‌‍‍‌‌‍‌‌‍‍‌‌‍‍​‍​‍​‍‍​‍​‍‌​‌‍​‌‌‍‍‌‍‍‌‌‌​‌‍‌​‍‍‌‍‍‌‌‍​‍​‍​‍​​‍​‍‌‍‍​‌​‍‌‍‌‌‌‍‌‍​‍​‍​‍‍​‍​‍‌‍‍​‌‌​‌‌​‌​​‌​​‍‍​‍​‍‌‍​‌‍‌‌​​‍‍‌​‌‌​‌‍​‌‌‍​‌‍‍‌‍‌‌‍‌‍‌‌‌​‍‌‍‌‍‌‍​‌‍‌‌​‍‍‌‍​‌‍​‍‌‍‍‌‌‍‍‌‌​‌‍‌‌‌‍‍‌‌​​‍‌‍‌‌‌‍‌​‌‍‍‌‌‌​​‍‌‍‌‌‍‌‍‌​‌‍‌‌​‌‌​​‌​‍‌‍‌‌‌​‌‍‌‌‌‍‍‌‌​‌‍​‌‌‌​‌‍‍‌‌‍‌‍‍​‍‌‍‍‌‌‍‌​​‌‌‍‌‍‌‍​‌​‍‌​​‌‌‍‌‍​‌‍‌‍​‌​‍​​‍‌​‌​‌‍​‍‌‍‌‍​​​​‍‌​‌​‌‍​‌​‌​‌‍​‌​‍‌‌‍​‌‌‍​​​​​‌​‍‌​‍​‌‍‌‌​‌‌‍‌‌‌‍‌‍‌‍‌​‌‍‌‌​‌​​‍​​​‌‌‌‍‌‍​‍‌‌​‌‍‌‌​​‌‍‌‌​‌‌‍​‍‌‍​‌‍‌‍‌‌‌​​‌‍‌​‌‌​​‍‌​​‌‍​‌‌‌​‌‍‍​​‌‌‌​‌‍‍‌‌‌​‌‍​‌‍‌‌​‌‍​‍‌‍​‌‌​‌‍‌‌‌‌‌‌‌​‍‌‍​​‌‌‍‍​‌‌​‌‌​‌​​‌​​‍‌‌​​‌​​‌​‍‌‌​​‍‌​‌‍​‍‌‌​​‍‌​‌‍‌‍​‌‍‌‌​​‍‍‌​‌‌​‌‍​‌‌‍​‌‍‍‌‍‌‌‍‌‍‌‌‌​‍‌‍‌‍‌‍​‌‍‌‌​‍‍‌‍​‌‍​‍‌‍‌‍‍‌‌‍‌​​‌‌‍‌‍‌‍​‌​‍‌​​‌‌‍‌‍​‌‍‌‍​‌​‍​​‍‌​‌​‌‍​‍‌‍‌‍​​​​‍‌​‌​‌‍​‌​‌​‌‍​‌​‍‌‌‍​‌‌‍​​​​​‌​‍‌​‍​‌‍‌‌​‌‌‍‌‌‌‍‌‍‌‍‌​‌‍‌‌​‌​​‍​​​‌‌‌‍‌‍​‍‌‍‌‌​‌‍‌‌​​‌‍‌‌​‌‌‍​‍‌‍​‌‍‌‍‌‌‌​​‌‍‌​‌‌​​‍‌‍‌​​‌‍​‌‌‌​‌‍‍​​‌‌‌​‌‍‍‌‌‌​‌‍​‌‍‌‌​‍‌‍‌​​‌‍‌‌‌​‍‌​‌​​‌‍‌‌‌‍​‌‌​‌‍‍‌‌‌‍‌‍‌‌​‌‌​​‌‌‌‌‍​‍‌‍​‌‍‍‌‌​‌‍‍​‌‍‌‌‌‍‌​​‍​‍‌‌ 4
When 2D Tasks Meet 1D Serialization: On Serialization Friction in Structured Tasks 4
Preference-Shaped Expected Hypervolume and R2 Improvement: Exact Computation and Monotonicity 4
Building Community-Centred NLP Resources for Puno Quechua 4
PrionNER: A Named Entity Recognition Dataset for Prion Disease Biomedical Literature 4
Breaking the Script Barrier: Enabling Automatic Alignment for PoS-based ASR Error Analysis in Non-Latin Scripts 4
The Cases LJP Never Sees: Prosecution Decision Prediction for More Complete Criminal Liability Assessment 4
GraphLit: Learning Text-Enriched Dynamic Character Network Representations for Literary Study 4
Agentic Separation Logic Specification Synthesis 4
Self-Consistency via Marginal Sharpening 4
Where Rollouts Begin: Low-Load, High-Leverage First-Token Diversification for RLVR 4
Entropy-aware Masking for Masked Language Modeling 4
Interpretability-Guided Layer Selection over Subspace Projection: SAEs as Stethoscopes, Not Scalpels, for Raw Task Vector Model Editing 4
Retention Consequence in Lifecycle Memory Control 4
Ultra-Reduced-Impact-Encased-Logging (URIEL): propose a new method for selective sustainable logging and post-harvest silvicultural treatment in tropical forest using airborne robotics systems 3
DELOS: Detecting Shallow Transits in Kepler Photometry Using a Contrastive-Learning Framework 3
v0.30.0-rc30 3
Final 24 hours to save up to $410 on your TechCrunch Disrupt 2026 ticket 3
Jony Ive’s funky Ferrari 3
Botnet of more than 17 million devices dismantled 3
The Download: unlocking lithium and controlling Ebola 3
The deadly Ebola outbreak is proving difficult to control 3
It's hard to justify buying a Framework 12 3
Function invocations now billed per unit 3
Hibernate 7.4 New Features 3
JetBrains Academy – May Digest 3
How Step Counters Work in Wearables and Why Different Devices Give Different Results 3
How Declarative Partial Updates Work in HTML 3
Best of the Heap: First post of the past​​​​‌‍​‍​‍‌‍‌​‍‌‍‍‌‌‍‌‌‍‍‌‌‍‍​‍​‍​‍‍​‍​‍‌​‌‍​‌‌‍‍‌‍‍‌‌‌​‌‍‌​‍‍‌‍‍‌‌‍​‍​‍​‍​​‍​‍‌‍‍​‌​‍‌‍‌‌‌‍‌‍​‍​‍​‍‍​‍​‍‌‍‍​‌‌​‌‌​‌​​‌​​‍‍​‍​‍‌‍​‌‍‌‌​​‍‍‌​‌‌​‌‍​‌‌‍​‌‍‍‌‍‌‌‍‌‍‌‌‌​‍‌‍‌‍‌‍​‌‍‌‌​‍‍‌‍​‌‍​‍‌‍‍‌‌‍‍‌‌​‌‍‌‌‌‍‍‌‌​​‍‌‍‌‌‌‍‌​‌‍‍‌‌‌​​‍‌‍‌‌‍‌‍‌​‌‍‌‌​‌‌​​‌​‍‌‍‌‌‌​‌‍‌‌‌‍‍‌‌​‌‍​‌‌‌​‌‍‍‌‌‍‌‍‍​‍‌‍‍‌‌‍‌​​‌​‌‍​‌‍​‍​‌‍‌‍​​‌​‌‍​​​‌‍​‍​‍‌‌‍​‍‌‍​​‌‌​‍​​‍‌​‌​​​‌‌‍‌‍​‌‌​‍‌‌‍​‍‌‍‌​​‍‌​​​​‍‌​‌​‌‌‌‍​‌​​​‍‌​​‌​​‍‌‍‌​​​‌​‌​​‌‌​‌‍​‍‌‌​‌‍‌‌​​‌‍‌‌​‌‌‍​‍‌‍​‌‍‌‍‌‌‌​​‌‍‌​‌‌​​‍‌​​‌‍​‌‌‌​‌‍‍​​‌‌‌​‌‍‍‌‌‌​‌‍​‌‍‌‌​‌‍​‍‌‍​‌‌​‌‍‌‌‌‌‌‌‌​‍‌‍​​‌‌‍‍​‌‌​‌‌​‌​​‌​​‍‌‌​​‌​​‌​‍‌‌​​‍‌​‌‍​‍‌‌​​‍‌​‌‍‌‍​‌‍‌‌​​‍‍‌​‌‌​‌‍​‌‌‍​‌‍‍‌‍‌‌‍‌‍‌‌‌​‍‌‍‌‍‌‍​‌‍‌‌​‍‍‌‍​‌‍​‍‌‍‌‍‍‌‌‍‌​​‌​‌‍​‌‍​‍​‌‍‌‍​​‌​‌‍​​​‌‍​‍​‍‌‌‍​‍‌‍​​‌‌​‍​​‍‌​‌​​​‌‌‍‌‍​‌‌​‍‌‌‍​‍‌‍‌​​‍‌​​​​‍‌​‌​‌‌‌‍​‌​​​‍‌​​‌​​‍‌‍‌​​​‌​‌​​‌‌​‌‍​‍‌‍‌‌​‌‍‌‌​​‌‍‌‌​‌‌‍​‍‌‍​‌‍‌‍‌‌‌​​‌‍‌​‌‌​​‍‌‍‌​​‌‍​‌‌‌​‌‍‍​​‌‌‌​‌‍‍‌‌‌​‌‍​‌‍‌‌​‍‌‍‌​​‌‍‌‌‌​‍‌​‌​​‌‍‌‌‌‍​‌‌​‌‍‍‌‌‌‍‌‍‌‌​‌‌​​‌‌‌‌‍​‍‌‍​‌‍‍‌‌​‌‍‍​‌‍‌‌‌‍‌​​‍​‍‌‌ 3
Online (one-pass) algorithms 3
Composer’s dependency policies 3
How to Build a PDF Page Numbering Tool in the Browser Using JavaScript 3
Comonadic Morphophonology: A Compositional Framework for Context-Dependent Morphological Rules in Finnish 3
One Group, Clearly, Is Deranged 2
Why people say CRTs don’t have pixels 2
DR DOS: Revenge of CP/M 2
This Week on The Analog Antiquarian 2
★ What Is a Dickover? 1
Xiaohu AI Daily · Curated AI Updates
2026-05-30 · Today: 800 picks