English(EN) LLMs Cosplay Understanding (Jeremy Howard)

Jeremy Howard：LLMs '模仿' 理解，而非真正领会

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-11 13:50

Jeremy Howard 认为，大型语言模型 (LLMs) 并不具备真正的理解能力，而是擅长模仿理解。他认为，它们出色的表现类似于角色扮演（cosplay），在没有真正内在推理的情况下，能够令人信服地模仿类似人类的反应。这一观点挑战了当前人工智能系统中涌现意识或理解能力的观念。 AI

影响挑战了对人工智能理解能力的认知，表明当前的 LLMs 是模仿而非理解。

排序理由由知名可信人士 (Jeremy Howard) 撰写的关于人工智能能力的观点文章。

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: Commentary; 由知名可信人士 (Jeremy Howard) 撰写的关于人工智能能力的观点文章。
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: opinion; 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: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

Machine Learning Street Talk TIER_1 English(EN) · Machine Learning Street Talk · 2026-11-11 13:50

大型语言模型扮演理解（Jeremy Howard）

Jeremy Howard, co-founder of fast.ai and Answer.AI, on the endless online argument over whether LLMs "really" understand. His answer is that both sides are right: LLMs cosplay understanding, much like the man in Searle's Chinese room. For a great many tasks the difference doesn't…