Blog
告别冗长 PDF，一文读懂最新顶会与核心期刊的创新价值。
Agentic Search in the Wild: Intents and Trajectory Dynamics from 14M+ Real Search Requests
Auto-ARGUE: LLM-Based Report Generation Evaluation
Emergent Coordination in Multi-Agent Language Models
A Reproducibility Study of LLM-Based Query Reformulation
Intern-Atlas: A Methodological Evolution Graph as Research Infrastructure for AI Scientists
Programming with Data: Test-Driven Data Engineering for Self-Improving LLMs from Raw Corpora
Targeted proteomics in a population-based study identifies serum PECAM-1 and TRIM21 as inflammation markers for periodontitis
ACLSum: A New Dataset for Aspect-based Summarization of Scientific Publications
Can Coding Agents Reproduce Findings in Computational Materials Science?
HVI-CIDNet+: Beyond Extreme Darkness for Low-Light Image Enhancement
H-RAG at SemEval-2026 Task 8: Hierarchical Parent-Child Retrieval for Multi-Turn RAG Conversations
SurGE: A Benchmark and Evaluation Framework for Scientific Survey Generation
Efficient Reasoning with Hidden Thinking
BoostDream: Efficient Refining for High-Quality Text-to-3D Generation from Multi-View Diffusion
Human-LLM Coevolution: Evidence from Academic Writing
Q-RAG: Long Context Multi-step Retrieval via Value-based Embedder Training
BoostDream: Efficient Refining for High-Quality Text-to-3D Generation from Multi-View Diffusion
Cognitive Alpha Mining via LLM-Driven Code-Based Evolution
Empirical Investigation of Latent Representational Dynamics in Large Language Models: A Manifold Evolution Perspective
Shadow in the Cache: Unveiling and Mitigating Privacy Risks of KV-cache in LLM Inference
FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness
A Theory of Generalization in Deep Learning
HeavySkill: Heavy Thinking as the Inner Skill in Agentic Harness
Safety of Semaglutide
Shape anisotropy governs organization of active rods: Swarming, turbulence, flocking, and jamming.
Can Coding Agents Reproduce Findings in Computational Materials Science?
OpenSeeker-v2: Pushing the Limits of Search Agents with Informative and High-Difficulty Trajectories
A Skill-Based AI Agentic Pipeline for Library of Congress Subject Indexing
HarmMetric Eval: Benchmarking Metrics and Judges for LLM Harmfulness Assessment
dots.ocr: Multilingual Document Layout Parsing in a Single Vision-Language Model