Neural Spectral Capacity: Measuring and Designing Architectures from Network Specification Alone Paper • 2609.23087 • Published Sep 19 • 9
DexJoCo: A Benchmark and Toolkit for Task-Oriented Dexterous Manipulation on MuJoCo Paper • 2605.16257 • Published May 15 • 55
Where Do Deep-Research Agents Go Wrong? Span-Level Error Localization in Agent Trajectories Paper • 2606.02060 • Published Jun 1 • 59