Dual Exam separates the evaluation of action selection from state evolution.
Dual Exam benchmarks agents and world models separately, exposing weak coupled performance
The authors test 10 agent configurations on 1,088 items and 16 world model configurations on 400 cases, finding top indices of 31.3 and 21.7 and zero successes in a 150-case coupled pilot.
Top University
Zhefan Rao · Meng Chu · Xuan Zhang · Jingfan Chen · Guohao Cai · Runyi Li · +77 more
The Hong Kong University of Science and Technology · National University of Singapore · The Hong Kong Polytechnic University · The University of Tokyo · Carnegie Mellon University
Research Digest··2 min read
The authors introduce Dual Exam, a two-track evaluation framework that tests multimodal agents under fixed environment responses and world models under fixed action interventions across physical, digital, social, and scientific domains.
§