Reliable confidence estimation is essential for trustworthy deployment of large language models (LLMs), but dedicated confidence training stage requires addi…
机构:UIUC
来源:arXiv 2610.31820 | AI4Papers 论文推荐平台