---
id: 20260912-T0-11
title: "NCP-ArchPreview：用下一概念预测把语言模型推进潜空间"
title_en: "NCP-ArchPreview Pushes Language Models into Latent Space via Next Concept Prediction"
url: https://ai.daily.yangsir.net/daily/20260912-T0-11
issue_date: 2026-09-12
publish_date: 2026-09-11T04:00:00.000Z
category: research
source_name: "arXiv cs.CL (NLP)"
source_url: https://arxiv.org/abs/2609.10715
---

# NCP-ArchPreview：用下一概念预测把语言模型推进潜空间

arXiv论文提出NCP-ArchPreview，一种潜空间语言模型，在标准下一词预测（NTP）之外引入下一概念预测（NCP），让模型学习预测跨区间的离散概念。该方法试图将自回归预训练从token层面推进到概念层面。目前为技术报告阶段，尚未公布具体基准测试结果。

## English Version

**NCP-ArchPreview Pushes Language Models into Latent Space via Next Concept Prediction**

An arXiv paper introduces NCP-ArchPreview, a latent-space language model that adds Next Concept Prediction (NCP) alongside standard next-token prediction (NTP), training the model to predict discrete concepts that span ranges. The approach pushes autoregressive pretraining from the token level toward the concept level. It is currently a technical report with no specific benchmark results disclosed yet.

---

**来源**：[arXiv cs.CL (NLP)](https://arxiv.org/abs/2609.10715)

**详情页**：https://ai.daily.yangsir.net/daily/20260912-T0-11

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*