3 papers
cs.LG2026
Learning Where Outcomes Change:Credit-Addressable Reasoning for Multimodal Geometry
Jiani Guo, Junjie Wang, Jie Wu +5
Multimodal geometry reasoning requires VLMs to extract precise visual relations and preserve them through multi-step deduction. Existing free-form traces obscure the decisions that…
physics.comp-ph2026
Cartesian tensor equivariant machine-learning force field for spin-dependent atomistic simulations
Junjie Wang, Yijie Zhu, Zhongwei Zhang +5
Magnetic materials exhibit an intricate coupling between atomic structure and spin degrees of freedom, posing a fundamental challenge for atomistic simulations across experimentall…
cs.CL2026
On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability
Zihan Qiu, Zekun Wang, Xiao Li +33
We describe the architecture and ablations of Qwen3.8-Flash-Next, a sparse mixture-of-experts model with 125B parameters, 6B activated per token, and additional 51B parameters of n…