arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2510.05069cs.CLcs.AI
收藏

SwiReasoning: 在潜在空间和显式空间中切换以实现帕累托更优推理的LLM

SwiReasoning: Switch-Thinking in Latent and Explicit for Pareto-Superior Reasoning LLMs

Georgia Tech(佐治亚理工学院);Microsoft(微软)

Dachuan Shi, Abedelkadir Asi, Keying Li, Xiangchi Yuan, Leyan Pan, Wenke Lee, Wen Xiao

AI总结:

SwiReasoning通过在潜在空间和显式空间之间切换,提升推理LLMs的准确率和令牌效率。

摘要:

最近的研究表明,除了通过显式步骤的离散推理之外,大型语言模型(LLMs)还能在潜在空间中进行连续推理,从而在每个步骤中提供更丰富的信息,从而提高令牌效率。尽管有这种前景,但潜在推理仍然面临两个挑战,尤其是在无训练设置中:1)纯粹的潜在推理通过维持多个隐式路径来扩展搜索分布,这会扩散概率质量,引入噪声,并阻碍收敛到单一的高置信度解决方案,从而损害准确性;2)即使没有显式文本,过度思考仍然存在,浪费令牌并降低效率。为了解决这些问题,我们引入了SwiReasoning,这是一个无训练的LLM推理框架,其两个关键创新是:1)SwiReasoning根据从下一个令牌分布中估计的熵趋势进行块级置信度,动态地在显式和潜在推理之间切换,以平衡探索和利用并促进及时收敛。2)通过限制最大思考块切换次数,SwiReasoning抑制过度思考并提高令牌效率,无论问题难度如何。在广泛使用的数学、STEM、编码和通用基准上,SwiReasoning在不同模型家族和规模的推理LLMs上一致提高了平均准确率1.8%-3.1%。此外,在受限预算下,SwiReasoning在令牌效率上提高了57%-79%,随着预算的收紧,收益更大。

英文摘要:

Recent work shows that, beyond discrete reasoning through explicit chain-of-thought steps, which are limited by the boundaries of natural languages, large language models (LLMs) can also reason continuously in latent space, allowing richer information per step and thereby improving token efficiency. Despite this promise, latent reasoning still faces two challenges, especially in training-free settings: 1) purely latent reasoning broadens the search distribution by maintaining multiple implicit paths, which diffuses probability mass, introduces noise, and impedes convergence to a single high-confidence solution, thereby hurting accuracy; and 2) overthinking persists even without explicit text, wasting tokens and degrading efficiency. To address these issues, we introduce SwiReasoning, a training-free framework for LLM reasoning which features two key innovations: 1) SwiReasoning dynamically switches between explicit and latent reasoning, guided by block-wise confidence estimated from entropy trends in next-token distributions, to balance exploration and exploitation and promote timely convergence. 2) By limiting the maximum number of thinking-block switches, SwiReasoning curbs overthinking and improves token efficiency across varying problem difficulties. On widely used mathematics, STEM, coding, and general benchmarks, SwiReasoning consistently improves average accuracy by 1.8%-3.1% across reasoning LLMs of different model families and scales. Furthermore, under constrained budgets, SwiReasoning improves average token efficiency by 57%-79%, with larger gains as budgets tighten.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑