LLMs Beat Humans in Fraud Detection & Investor Pressure
Large language models outperform humans in detecting investment fraud and resisting investor pressure, enhancing financial advisory reliability.
CHORUS: AI Framework for Realistic Online Deliberation Data
Discover CHORUS, an AI framework generating realistic online deliberation data with agentic actors and temporal modeling for authentic discussions.
pAI/MSc: Human-in-the-Loop ML Research Tool
Discover pAI/MSc, an open-source platform enhancing ML theory research with a human-in-the-loop approach for efficient academic workflows.
Self-Guided Plan Extraction with Goal-Conditional RL
Discover SuperIgor, a framework using self-guided plan extraction and goal-conditional reinforcement learning for adaptive instruction-following AI.
Evaluating Generative AI as Pluralist Sociotechnical Systems
Explore a pluralist framework for evaluating generative AI, focusing on sociotechnical interactions, values, and responsible benchmarks beyond static measu...