Category: Large Language Models

• Lin Junyang Speaks Out for the First Time After Leaving Alibaba: Reviewing Qwen's Detours, Pointing to AI's New Path
• VideoSeek Long-Video Understanding Agent: The Secret to Boosting GPT-5's Long-Video Comprehension by 10 Points
• Let AI 'Refine' Its Own Data! DataChef Goes Open Source: Using Reinforcement Learning to Automatically Generate LLM Data Recipes
• The AI Grad Student Who Can Do Theoretical Physics Is Here—Is There No Going Back? A Harvard Professor's Insights You Must Read
• TurboQuant: Redefining AI efficiency with extreme compression
• NVIDIA Nemotron-Cascade 2 Technical Report
• ICLR 2026 | How Far Can Unsupervised Reinforcement Learning Go for Large Models? A Systematic Answer from the Tsinghua Team
• Stop Obsessing Over Outcome Rewards! CUHK Identifies and Solves the "Information Self-Locking" Problem in RL!
• One Year into RAG: My Biggest Regret Was Adopting Knowledge Graphs
• Karpathy Just Open-Sourced AutoResearch: I Used It to Optimize Lobster Skills, Boosting Success Rates from 56% to 92%
• The Age of Agent Skills: How Big is the Gap Between Strong and Weak Models? Shattering the 'Cheap Alternative' Illusion | Latest from Oxford
• OpenAI's New Model Rejected on Day 0! Ranks Poorly, Lagging Behind Domestic Models Released in Late January
• 4B Model Surpasses GPT-5 in Hallucination Suppression: CMU and Others Propose New Behaviorally Calibrated Reinforcement Learning Method
• Masterpiece! MIT and Google Train an LLM Capable of Rigorous Bayesian Inference
• Breaking Static Model Weights! Tencent Hunyuan Releases Real-Time Brain-Switching Technology for Inference
• MMLU is Dead? 'Humanity's Last Exam' Published in Nature: Global AI Models Collectively Fail!
• Why Can Large Language Models 'Understand' the World?
• Google's New Research Identifies the Crucial Tokens Where Large Models Ponder Deeply!
• Anthropic CEO: The Data Bottleneck for Large Models No Longer Exists, Models Are Training Themselves
• Anthropic's Latest Paper: Internet Anonymity Ends in the AI Era | Hao's Paper Chat
• Not All Tokens Are Equal! Google Proposes True Deep Thinking: Long Chain of Thought ≠ Deep Reasoning
• Qwen3.5-Flash Arrives! Three Medium-Scale Models Open-Sourced
• Are LLM RL Training Trajectories Actually Linear? Miaow Lab's Latest Work: Directly 'Predict' Future Models Without Further Training!
• What Exactly Is On-Policy Distillation? An In-Depth Interpretation of On-Policy/Self-Distillation
• Behind the $1 Trillion Evaporation: The Moats of Vertical Software Are Being Rewritten by Large Models