Authors: || Published: 2025-07-26T12:32:00 || Updated: 2025-07-26T12:32:00 || 2 min read
Categories: || Tags: || Post-format: link
Recent AI Reading [26 July 2025]
Papers
Agentic AI / AI Agents
- Mind2Web 2: Evaluating Agentic Search with Agent-as-a-Judge
- Establishing Best Practices for Building Rigorous Agentic Benchmarks
- Survey on Evaluation of LLM-based Agents
Large Language Models
- Learning to Think: Information-Theoretic Reinforcement Fine-Tuning for LLMs
- Reinforcing Multi-Turn Reasoning in LLM Agents via Turn-Level Credit Assignment
- The Ultimate Guide to Fine-Tuning LLMs from Basics to Breakthroughs: An Exhaustive Review of Technologies, Research, Best Practices, Applied Research Challenges and Opportunities
- THiNK: Can Large Language Models Think-aloud?
- Transformers are Graph Neural Networks
- REST: Stress Testing Large Reasoning Models by Asking Multiple Problems at Once
- A Practical Two-Stage Recipe for Mathematical LLMs: Maximizing Accuracy with SFT and Efficiency with Reinforcement Learning
- A Survey of Context Engineering for Large Language Models
- Training Large Language Models to Reason in a Continuous Latent Space
- Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety
- Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task
- Responsible Artificial Intelligence Systems: A Roadmap to Society’s Trust through Trustworthy AI, Auditability, Accountability, and Governance
- Distribution of responsibility for AI development: expert views
- Identifying AI Hazards and Responsibility Gaps
- Thinking Beyond Tokens: From Brain-Inspired Intelligence to Cognitive Foundations for Artificial General Intelligence and its Societal Impact
Articles and Blog Posts
- How we built our multi-agent research system
- WEF Report: The Impact of AI Driving 170m New Jobs by 2030
- Is there a Half-Life for the Success Rates of AI Agents?
- Scaling Reinforcement Learning: Environments, Reward Hacking, Agents, Scaling Data
- The Path to Medical Superintelligence
- Kimi K2: Open Agentic Intelligence
- Trends – Artificial Intelligence
- There Are No New Ideas in AI… Only New Datasets
- Reasoning + RL: A new recipe for AI apps