Large language models (LLMs) are increasingly deployed to solve complex scientific and practical problems via iterative optimization. However, dynamically co…
机构:字节
来源:arXiv 2609.37221 | AI4Papers 论文推荐平台