Parallel thinking has emerged as a promising paradigm for enhancing the reasoning capabilities of Large Language Models (LLMs) by exploring multiple reasonin…
机构:北大
来源:arXiv 2610.29501 | AI4Papers 论文推荐平台