1 paper · 1 filter
Cheng Yan, Guangyang Ye, Wuyang Zhang +5
While test-time scaling improves the problem-solving ability of large reasoning models (LRMs) through additional inference-time computation, it can also exacerbate overthinking and…