1 paper · 1 filter
Wei Li, Yanbin Wei, Qiushi Huang +4
Modern large language models (LLMs) often struggle to dynamically adapt their reasoning depth to varying task complexities, leading to suboptimal performance or inefficient resourc…