PulseAugur
中
实时 10:28:54
English(EN) Learning More by Training Less: Uncertainty-Guided Progressive Contraction for Efficient Reasoning Post-Training

新研究提供高效的 LLM 推理和遗忘方法 · 已追踪 2 个来源

两篇新研究论文提出了使大型语言模型在推理能力方面更高效的方法。第一篇“通过更少的训练学到更多”(Learning More by Training Less)介绍了一种不确定性引导的渐进收缩框架,通过将优化集中在关键的推理轨迹上,显著降低了监督微调所需的训练预算。第二篇论文“REDaCT”提出了一种推理时遗忘框架,旨在从推理模型中移除特定信息,而不会导致性能下降或“推理崩溃”(模型无法产生连贯思维过程的常见问题)。 AI

影响 这些方法可能带来更高效的训练和对 LLM 输出的更好控制,从而可能降低计算成本并提高模型可靠性。

排序理由 该集群包含两篇详细介绍改进大型语言模型新方法的学术论文,符合研究类别。

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →

新研究提供高效的 LLM 推理和遗忘方法 · 已追踪 2 个来源

本文如何被排名

Signal score
17 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
该集群包含两篇详细介绍改进大型语言模型新方法的学术论文,符合研究类别。
Source corroboration
2 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
paper, other
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 [2]

  1. arXiv cs.AI TIER_1 English(EN) · Suorong Yang, Jiaxin Wang, Jiaming Su, Chentao Yue, Furao Shen, Peng Ye, Soujanya Poria ·

    训练更少,学习更多:不确定性引导的渐进收缩,用于高效的训练后推理

    arXiv:2610.12933v1 Announce Type: new Abstract: Large language models have achieved remarkable advances in complex reasoning. However, scaling reasoning post-training with ever-larger corpora can substantially increase training cost without proportional performance gains. In this…

  2. arXiv cs.AI TIER_1 English(EN) · Joshua Maccoby, Wendy Hui Wang ·

    REDaCT:大型推理模型中的无崩溃解学

    arXiv:2610.12769v1 Announce Type: cross Abstract: Unlearning in large reasoning models (LRMs) requires removing forget-set information from both the reasoning trace and the final answer. Existing approaches achieve this objective at the cost of reasoning collapse, where the model…