English(EN) Do Latent Tokens Think? A Causal and Adversarial Analysis of Chain-of-Continuous-Thought

新研究质疑潜在标记在大型语言模型推理中的作用

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

一篇来自arXiv的新论文对大型语言模型中“潜在标记”的有效性提出了质疑，特别是在连续思想链（COCONUT）框架内。研究人员Yuyi Zhang及其同事发现，这些潜在标记充当了不可解释的占位符，而不是编码真正的推理过程。他们的实验表明，COCONUT虽然看起来稳定，但主要利用了数据集的伪影和捷径，导致在MMLU和HotpotQA等基准测试上的表现虚高，而未能展示真正的推理能力。 AI

影响挑战了某些大型语言模型推理技术的可靠性，可能影响未来的模型开发和评估方法。

排序理由在arXiv上发表的学术论文，分析了大型语言模型的推理机制。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 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; 在arXiv上发表的学术论文，分析了大型语言模型的推理机制。[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, safety; 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.AI TIER_1 English(EN) · Yuyi Zhang, Boyu Tang, Tianjie Ju, Sufeng Duan, Gongshen Liu · 2026-10-27 04:00

潜在标记会思考吗？对连续思维链的因果和对抗性分析

arXiv:2512.21711v2 Announce Type: replace-cross Abstract: Latent tokens are gaining attention for enhancing reasoning in large language models (LLMs), yet their internal mechanisms remain unclear. This paper examines the problem from a reliability perspective, uncovering fundamen…

本文如何被排名

报道来源 [1]

潜在标记会思考吗？对连续思维链的因果和对抗性分析

相关实体

相关话题