English(EN) Two For The Price Of One: Curvature-Aware Pruning Preserves Influence Functions Too

新的CACTUS优化器通过保留影响函数来增强LLM剪枝

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-14 04:00

研究人员开发了CACTUS，一种新颖的优化器，它使用感知曲率的目标来改进大型语言模型的训练后剪枝。该方法旨在保留稀疏模型和密集模型之间的内部梯度对齐，这是传统剪枝技术通常缺乏的特性。CACTUS在LLM-Surgeon等现有方法上展示了显著的加速，并在用于数据选择时，在BIG-Bench Hard等基准测试中表现出改进的性能。 AI

影响这项研究可能通过实现显著的剪枝而不会造成实质性的性能下降，从而导致更高效的大型语言模型部署。

排序理由该集群包含一篇详细介绍模型剪枝新方法的论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv stat.ML 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: 3 / 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模型剪枝新方法的论文。[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 stat.ML TIER_1 English(EN) · Mehdi Makni, Natalia Ponomareva, Rahul Mazumder · 2026-10-14 04:00

一石二鸟：感知曲率的剪枝也能保留影响函数

arXiv:2610.18516v1 Announce Type: cross Abstract: Post-training pruning enables efficient deployment, but conventional methods do not explicitly preserve internal-gradient alignment between sparse and dense models. We study novel connections between downstream performance of comp…

本文如何被排名

报道来源 [1]

一石二鸟：感知曲率的剪枝也能保留影响函数

相关实体

相关话题