DREAM: Deep Research Evaluation with Agentic Metrics
• DREAM introduces agentic evaluation for AI research agents, addressing lack of ground truth. • Highlights Mirage of Synthesis: surface fluency can mask factual and reasoning flaw
Agenticmetrics