Aligning LLM Agents with Human Social Values Using GraphRAG
Discover how GraphRAG enhances LLM-based agents to align with human social values, improving AI decision-making and emotional understanding.
Detecting Scientific Theory Shifts in AI with Sheaf Theory
Explore a sheaf-theoretic framework to detect and manage scientific theory shifts in AI agents, improving adaptability and coherence in AI research.
Conditional Attribute Estimation with Autoregressive Models
Discover how Conditional Attribute Transformers improve sequence modeling by enhancing attribute estimation and steerable generation in autoregressive mode...
PolitNuggets: Benchmarking AI Discovery of Political Facts
PolitNuggets benchmark enhances AI's ability to discover and synthesize long-tail political facts across languages with advanced multi-agent evaluation.
Preping: Efficient Agent Memory Building Without Tasks
Discover Preping, a novel AI framework enabling agent memory construction through self-generated tasks, reducing cold-start gaps and boosting efficiency.