LLM Reasoning Reveals Myopic Planning in Search Trees
Study uncovers shallow search depth and myopic planning in LLMs, highlighting key differences from human decision-making strategies.
Structured Randomness Boosts Multi-Agent Coordination
Discover how structured randomness enhances coordination in multi-agent systems, enabling flexible and scalable cooperation across tasks.
Uneven Cognitive Growth in Generative AI Models Over Time
Explore how generative AI models show uneven cognitive evolution, excelling in language but lagging in perceptual reasoning across generations.
Agent-BOM: Unified Security Auditing for LLM Agents
Discover Agent-BOM, a novel graph-based framework enhancing security audits for LLM agentic systems by bridging semantic gaps and assessing risks.
SCALAR: Enhancing AI Reasoning in Theoretical Physics
Discover how SCALAR's Actor-Critic-Judge loop improves AI-assisted theoretical physics through optimized feedback and multi-turn dialogues.