As Large Language Models (LLMs) increasingly serve as foundational reasoning engines, their tendency to hallucinate remains a critical vulnerability. While r…
机构:南洋理工大学
来源:arXiv 2610.02066 | AI4Papers 论文推荐平台