Blog
No more endless PDFs. Discover the core value of the latest top-tier research in one article.
AQuA: Recursively Self-Improving Quantitative Trading Research Agents
AgentFold: Closed-Loop Agentic Search for Protein Folding Model Design
One Token Is Enough: Fingerprinting and Verifying Large Language Models from Single-Token Output Distributions
Why Do AI Agents Break Rules? How Framing, Context, and Social Signals Shape Compliance
Language Models Can Control Their Own Attention
Harness-of-Harness: Multi-Day Autonomous Software Development with Continual Improvement
WHALE: A Simple Recipe for Joint Harness-Weight Optimization
Almost Free State Prediction Separation
STAIR (STructure Aware Information Retriever): A novel dataset and LLM based retriever for document structure augmentation
Large-Language Models as a Cognitive Virus
9.8:2
MoBA: Mixture of Block Attention for Long-Context LLMs.
9.9:2
9.11:1
9.11:2
The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement
World in World: Explore the World with World Models
GAUGE: When Not to Trust LLM-as-a-Judge in User-Simulated Evaluation of Task-Oriented Agents
Theory is all you need: AI, human cognition, and causal reasoning
Conceptual innovation and practical value of China's citywide industry-education consortium in vocational education