zip2zip++ 方法在最小化质量损失的情况下压缩 LLM 令牌

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-12 04:00

研究人员开发了 zip2zip++，一种在推理时压缩语言模型令牌且质量损失极小的方法。该技术解决了嵌入冲突、合并令牌的 RoPE 应用不正确以及因合并数字导致的准确性下降等问题。通过解耦嵌入、修改 RoPE 定位以及增强 LZW 编解码器，zip2zip++ 显著提高了困惑度（perplexity）和数学推理准确性，缩小了与未压缩模型相比的质量差距。 AI

影响这项研究可能通过在推理过程中减少令牌使用而不牺牲性能，从而实现更高效的语言模型。

排序理由该条目是一篇学术论文，详细介绍了一种新的 LLM 令牌压缩方法。[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: 11 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该条目是一篇学术论文，详细介绍了一种新的 LLM 令牌压缩方法。[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, infra; 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) · Saibo Geng, Andrea Gentilini, Xinxian Ma, Nathan Ranchin, Yunzhen Yao, Robert West · 2026-10-12 04:00

zip2zip++：迈向推理时令牌压缩，实现最小质量损失

arXiv:2610.13605v1 Announce Type: new Abstract: Dynamic vocabularies built on-the-fly via Lempel-Ziv-Welch (LZW) compression let a language model represent the same text in fewer tokens, but prior work shows that this efficiency comes at a cost to reasoning and perplexity. We sho…

本文如何被排名

报道来源 [1]

zip2zip++：迈向推理时令牌压缩，实现最小质量损失

相关实体

相关话题