Category: Large Language Models

• More Capable Than Gemini Diffusion! The First Multimodal Large Diffusion Language Model MMaDA Released, Achieving Strong Reasoning and High Controllability
• Does AI Know When to "Think"? Thinkless Teaches Large Language Models When to Reason
• ICML 2025 | Training-Free, Instant Alignment of Large Model Preferences
• Google | Tracing RAG System Errors: Proposing a Selective Generation Framework to Boost RAG Accuracy by 10%
• Multimodal Large Models Collectively Fail, GPT-4o Only 50% Safety Pass Rate: SIUO Reveals Cross-Modal Safety Blind Spots
• Nature Sub-journal: Humans Lost to AI Again, Especially When It Knows Who You Are
• Which Model Should a Reliable Agent Use? The "Lost in Conversation" Phenomenon in LLM Multi-turn Dialogues | Microsoft Latest
• When Thinking Becomes a Burden: Unveiling the "Thinking Traps" of Large Language Models
• How Strong is the Reasoning Ability of Large Language Models? A Study Reveals LLMs' Limitations and Potential
• Breakthrough in Reasoning: How SoftCoT++ Enables LLMs to 'Think Multiple Paths'?
• Qwen Breakthrough: Using "Parallel Computing" Instead of "Stacking Parameters", New Method Reduces Memory by 22x, Latency by 6x
• ZeroSearch: <Alibaba Technology> Large Language Models Learn Through Self-Rewarding Without a Browser
• Jeff Dean: AI Will Replace Junior Engineers Within a Year, Netizens: "Altman Only Pitches, What Jeff Says Is Fatal"
• AM-Thinking-v1: Advancing the Frontier of Reasoning at 32B Scale
• Ant Group's Wu Wei: A Big Guess on the Next Generation 'Reasoning' Model Paradigm
• From Intuition to "Deep Thinking": Multidimensional Evolution of Large Model Reasoning Capabilities
• DeepSeek Accuracy and Efficiency Doubled, Huawei & CAS Propose Chain-of-Thought "Early Exit" Mechanism
• GPT-5 R&D Insider Details Revealed! OpenAI Chief Research Officer: AGI is Just Around the Corner
• ZeroSearch: Zero-Search Reinforcement Incentivizes Model Potential, Ushering in a New Era for LLM Search Capability
• Forcing Models to Argue with Themselves, Recursive Thinking CoT Version Soars in Popularity! Netizens: Isn't This Just the Usual Trick for Most Reasoning Models?
• Stanford's Weak-for-Strong (W4S): Harnessing Stronger LLMs with Meta-Agent, Accuracy Boosted to 95.4% | Latest
• Can a single data point significantly enhance the mathematical reasoning performance of large models?
• Research: LLM's Prefilling Feature Has Become Its Jailbreak Vulnerability!
• PKU, Tsinghua, UvA, CMU, etc. Jointly Release: Latest Survey on Logical Reasoning Abilities of Large Models
• NVIDIA's Llama Nemotron Series: Key Technologies Explained