1 paper
Leheng Sheng, An Zhang, Zijian Wu +5
Recent studies empirically reveal that large reasoning models (LRMs) can automatically allocate more reasoning strengths (i.e., the number of reasoning tokens) for harder problems,…