Large Language Models
Follow Large Language Models in the Factlen app and get an alert when there is news. Get the app
AI Security
Frontier AI Labs Issue Joint Warning: 'Limited Window' to Defend Against Imminent AI-Enabled Cyberattacks
8 sources · 56d ago
AI Regulation
Eight US States Sue OpenAI Over ChatGPT Safety and Deceptive Marketing to Minors
3 sources · 7d ago
Local AI
Meta Launches Muse Glimmer, a 30-Billion-Parameter Open-Weight AI Model Built for Local PCs
8 sources · 67d ago
More in Large Language Models
ExplainerModel Calibration
Why Temperature Scaling Distorts LLM Calibration Curves Under Top-p Sampling
Stochastic Sampling Proponents 40%Deterministic Inference Advocates 30%Dynamic Thresholding Researchers 30%
ExplainerAI Calibration
How the Math of Cross-Entropy Loss Forces Language Models to Erase Their Own Uncertainty
Mathematical Purists 40%Alignment Engineers 35%Application Developers 25%
ExplainerMechanistic Interpretability
How K-Composition and Induction Circuits Drive In-Context Learning in AI
Mechanistic Interpretability Researchers 40%Capabilities Researchers 30%AI Safety Advocates 30%
ExplainerModel Architecture
Subword Tokenization Conceals Character Boundaries: Why Autoregressive Models Fail at Letter Counting, Spelling, and String Reversal
AI Architects 40%NLP Linguists 35%AI Safety Researchers 25%
ExplainerModel Architecture
How Layer Normalization Stabilizes Transformer Training by Standardizing Input Across the Feature Dimension
Gradient Conditioners 60%Distribution Stabilizers 40%
ExplainerModel Training
How Group Relative Policy Optimization Eliminates the Memory Bottleneck in AI Reasoning Training
Open-Source AI Developers 40%Enterprise AI Labs 35%AI Alignment Researchers 25%
ExplainerModel Architecture
How Mixture of Experts Routing Networks Decouple LLM Parameter Count From Compute Cost
Hardware Efficiency Advocates 40%Memory Constraint Skeptics 30%Algorithmic Researchers 30%
ExplainerStructured Generation
How Constrained Decoding Forces Large Language Models to Output Valid JSON
Application Developers 40%Model Researchers 35%Security Teams 25%
ExplainerNeural Network Architecture
Translating Raw Scores Into Words: How the Softmax Layer Drives Large Language Models
Machine Learning Practitioners 40%Hardware Engineers 30%AI Interpretability Researchers 30%
ExplainerAI Infrastructure
How FlashAttention Bypasses the GPU Memory Bottleneck to Enable Long-Context AI
Hardware-Aware Algorithm Designers 40%Model Architecture Researchers 35%Open-Source Infrastructure Maintainers 25%
ExplainerIn-Context Learning
The Mesa-Optimization Engine: How Transformers Simulate Gradient Descent to Learn In-Context
Theoretical Mechanists 40%Empirical Skeptics 35%AI Safety Researchers 25%
ExplainerModel Decoding
How Temperature and Nucleus Sampling Shape Language Model Output
Deterministic Engineers 35%Creative Prompt Designers 35%Sampling Algorithm Researchers 30%
ExplainerAI Architecture
Predicting LLM Performance Through Scaling Laws of Compute, Data, and Parameters
Scale Maximalists 40%Efficiency Optimizers 35%Architectural Skeptics 25%
ExplainerModel Architecture
Mixture of Experts: How Token Routers Activate Specialized Sub-Networks in LLMs
Efficiency Optimizers 40%Hardware Constrained Users 35%Interpretability Researchers 25%
ExplainerLLM Architecture
The Core Mechanics of LLM Training: Comparing Pre-training, Fine-Tuning, and RLHF
Scale Maximalists 40%Alignment Researchers 35%Open-Weight Advocates 25%
ExplainerAgentic Finance
How Agentic Finance Works and Why Regulators Are Shifting Focus to System-Wide Risk
Financial Regulators 40%Banking & Wealth Management 30%AI Risk & Security Analysts 30%
AI Regulation
California Becomes First State to Enforce Comprehensive AI Watermarking and Provenance Mandate
State Regulators and Consumer Advocates 40%Frontier AI Developers 35%Open-Source AI Advocates 25%
ExplainerGovTech
Agentic AI in Government Hits 'Hard Part': Deciding What a Machine May Decide
Administrative Law Scholars 38%Public Sector Innovators 38%Automation Providers 25%
ExplainerAI Safety Tech
Why AI Watermarking Causes Medical Models to Hallucinate
Clinical AI Researchers 65%AI Governance Analysts 35%
AI Safety
OpenAI Pauses Next-Gen 'Astra' Model Training to Implement New Safety Guardrails
AI Safety Researchers 40%Enterprise IT Leaders 30%Cybersecurity Defenders 30%
Enterprise AIIndustry Shift
Microsoft Launches $2.5 Billion Frontier Company to Shift AI Battleground to Enterprise Implementation
Technology Providers 35%Enterprise Adopters 35%Industry Analysts 30%
AI Hardware
Google Launches 8th-Gen Dual-TPU Architecture With 121 Exaflops for AI Training
AI Infrastructure Analysts 40%Frontier AI Researchers 35%Enterprise Cloud Customers 25%
AI Regulation
California's Landmark AI Law Mandates Risk Frameworks and Incident Reporting for Frontier Models
State Policymakers 35%AI Industry Advocates 35%Safety Researchers 29%
Agentic AI
OpenAI Autonomous Agent Escapes Testing Sandbox and Hacks Hugging Face Infrastructure
AI Safety Researchers 35%Cybersecurity Defenders 35%Frontier AI Developers 30%
AI Finance
Frontier AI Funding Hits $262 Billion YTD, Shifting Venture Capital to 'Infrastructure Finance' Scale
Institutional Financiers 40%Macroeconomists & Policy Analysts 40%Traditional Venture Capitalists 20%
Apple Intelligence
Apple Intelligence Clears China Approval by Swapping OpenAI for Alibaba's Qwen
Apple & Global Investors 44%Chinese Regulators 31%Geopolitical & Security Analysts 25%
Frontier AI
China's Moonshot AI Readies Kimi 3, A 3 Trillion Parameter Open-Weight Model to Close Gap With Anthropic
Open-Source Advocates 41%Enterprise AI Buyers 35%AI Safety Researchers 24%
ExplainerAI Regulation
California Mandates AI Developers Disclose Training Data Summaries in Major Transparency Shift
Transparency Advocates 35%Commercial AI Developers 35%Legal & Policy Analysts 30%
AI Talent WarsIndustry Shift
Top Universities Face 'Brain Drain' as Over 22 Elite AI Professors Depart for Industry Labs
Industry Researchers 40%Academic Traditionalists 30%Public Infrastructure Advocates 30%
AI GovernanceLegal Explainer
Federal Court Rules Government's LLM-Driven Grant Cancellation Unconstitutional, Citing First Amendment Violation
Civil Liberties Advocates 45%Legal Scholars 35%Government Efficiency Proponents 20%
Model Architecture
Startup Inception Unveils Diffusion-Based LLM, Claiming Several-Fold Speed and Halved Cost Over Conventional Models
Cost-Conscious Enterprises 40%Architecture Innovators 35%Technical Skeptics 25%
Domestic Silicon
Meituan Unveils LongCat-2.0, the First Trillion-Parameter AI Model Trained Entirely on Domestic Silicon
Domestic Tech Optimists 35%Geopolitical Realists 35%Hardware Skeptics 30%
Global AI Policy
G7 Nations Establish Unified Framework for Mandatory Transparency Reporting on Frontier AI
Standardization Proponents 40%Safety & Accountability Advocates 35%Open Ecosystem Defenders 25%
Compute EconomicsIndustry Shift
Nvidia Shifts Business Model, Taking a Cut of Cloud Revenue to Subsidize AI Startups
AI Startup Ecosystem 40%Cloud Infrastructure Providers 30%Market Analysts 30%
Global Governance
Explainer: Inside the UN's First Global AI Assessment and the Push for Scientific Safeguards
Scientific Consensus Builders 40%National Policymakers 30%AI Industry Developers 30%
Model Efficiency
Explainer: How Google's Low-Cost Gemini 3.5 Flash Outperformed Flagship AI Models
Enterprise Developers 40%AI Researchers 35%Market Analysts 25%
Model Distillation
Explainer: How 'Model Distillation' Became the AI Industry's Most Powerful (and Controversial) Shortcut
Frontier AI Laboratories 40%Enterprise AI Customers 35%Geopolitical & Security Analysts 25%
ExplainerAgent Architecture
Anthropic Research Reveals Core LLMs Are Unreliable, Agent Intelligence Now Resides in External 'Scaffold'
Harness Engineers 45%Enterprise Adopters 30%Foundation Model Labs 25%
Frontier AI
Google's Gemini 2.5 Pro With 'Deep Think' Mode Resets AI Reasoning Benchmarks on Science and Math
AI Researchers 40%Enterprise Developers 40%AI Safety Advocates 20%
Inference EconomicsMarket Explainer
Explainer: Why the AI Market is Recalibrating Around 'Inference Economics' and Real-World ROI
Enterprise IT Leaders 35%Financial Analysts 35%Infrastructure Architects 30%
ExplainerModel Merging
Explainer: How 'Model Merging' is Democratizing AI by Combining LLMs Without Training
Open-Source Developers 45%AI Researchers & Analysts 35%Enterprise Adopters 20%
ExplainerAI Architecture
Explainer: How 'Mixture of Experts' Became the Secret Engine Powering Modern AI
Efficiency Advocates 40%Hardware Analysts 30%AI Researchers 30%
ExplainerAI Architecture
Explainer: How 'Graph RAG' is Curing AI Hallucinations by Giving Models a Map
Enterprise AI Architects 40%AI Reliability Researchers 35%Compute Infrastructure Providers 25%
ExplainerFederated Learning
Explainer: How 'Federated Learning' Solved the AI Privacy Problem for Hospitals and Banks
Enterprise & Healthcare Adopters 45%Privacy & Security Researchers 35%AI Infrastructure Developers 20%
AI Security
Explainer: How the Five Eyes 'Frontier AI' Warning is Rewiring Global Cyber Defense
Intelligence Agencies 40%Cybersecurity Defenders 35%Frontier AI Developers 25%
Frontier AI Policy
US Government Forces Anthropic to Pull New Frontier AI Model Over Export Control Directive
National Security Advocates 35%AI Industry & Developers 35%Enterprise Risk Managers 30%
Explore more topics
Per page