DeepResearchEval: An Automated Framework for Deep Research Task Construction and Agentic Evaluation
IntermediateYibo Wang, Lei Wang et al.Jan 14arXiv
The paper introduces DeepResearchEval, a fully automated way to build realistic deep research tasks and to grade long research reports from AI systems.
#deep research agents#agentic evaluation#persona-driven tasks