Large reasoning models achieve strong performance on challenging tasks, but their long chain-of-thought often incurs substantial inference costs by continuin…
机构：西安交大
来源：arXiv 2611.04814 | AI4Papers 论文推荐平台