English(EN) One Model, Two Paths: Balancing Single-Pass and Looped Performance in Language Models

新方法平衡语言模型的单次推理和循环推理

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-27 04:00

研究人员开发了一种方法，通过重用中间层权重来使预训练语言模型适应单次推理和循环推理。这种方法通过增加计算量来提高性能，而无需增加参数。然而，为循环预测而进行的训练可能会对单次推理质量产生负面影响。该研究提出了联合训练策略，使用加权损失和来自冻结的原始模型的指导来减轻这种退化，同时保留循环推理的优势。 AI

影响这项研究可能通过使单个模型在单次推理和循环推理模式下都能表现良好，从而提高语言模型推理的效率。

排序理由该集群包含一篇详细介绍语言模型新适应方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 10 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍语言模型新适应方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CL TIER_1 English(EN) · Jiarun Shen, Lei Zhao, Xiao Li · 2026-10-27 04:00

单一模型，两条路径：平衡语言模型的单次遍历与循环性能

arXiv:2610.33517v1 Announce Type: new Abstract: Repeating Transformer layers can improve language-model performance without adding parameters. A pretrained model can be adapted to support both single-pass and looped inference by reusing middle-layer weights for additional computa…

本文如何被排名

报道来源 [1]

单一模型，两条路径：平衡语言模型的单次遍历与循环性能

相关实体

相关话题