Lingjiao Chen

2025年4月15日

Large language models (LLMs) are increasingly capable of complex reasoning through “inference-time scaling,” a set of techniques that allocate more computational resources during inference to generate answers. However, a new study from Microsoft Research reveals that the effectiveness of these…