[人人能懂] 从内功心法、内心戏到学习公式 episode artwork

EPISODE · Oct 31, 2025 · 27 MIN

[人人能懂] 从内功心法、内心戏到学习公式

from AI可可AI生活

本期节目,我们将一起深入AI的大脑,看看它究竟是如何思考的。我们会发现,AI不仅会修炼“内功心法”,以小博大;还会上演复杂的“内心戏”,让我们难辨其思考的真伪。我们还会揭开它通用的“学习公式”,看看聪明的AI为何会掉入“聪明陷阱”,以及它最终如何学会替我们高效地“试错”。00:00:32 AI的“内功心法”:让小模型拥有大智慧的秘密00:05:52 AI的“内心戏”:你看到的思考过程,有多少是“演”的?00:10:49 AI的“聪明陷阱”:为什么懂得多,反而容易犯错?00:16:26 揭秘AI的“学习公式”:原来万变不离其宗00:21:32 让AI替你“试错”,我们能省下多少力气?[CL] Scaling Latent Reasoning via Looped Language Models[ByteDance Seed]https://arxiv.org/abs/2510.25741---[LG] Can Aha Moments Be Fake? Identifying True and Decorative Thinking Steps in Chain-of-Thought[Northeastern University & UC Berkeley]https://arxiv.org/abs/2510.24941---[CL] Are Language Models Efficient Reasoners? A Perspective from Logic Programming[ETH Zürich & EPFL]https://arxiv.org/abs/2510.25626---[CL] Language Model Behavioral Phases are Consistent Across Architecture, Training Data, and Scale[MIT & UCSD]https://arxiv.org/abs/2510.24963---[LG] GPTOpt: Towards Efficient LLM-Based Black-Box Optimization[MIT]https://arxiv.org/abs/2510.25404在小宇宙查看该单集文稿

Episode metadata supplied by the publisher feed · Published Oct 31, 2025

Embed this episode

NOW PLAYING

[人人能懂] 从内功心法、内心戏到学习公式

0:00 27:07

No transcript for this episode yet

We transcribe on demand. Request one and we'll notify you when it's ready — usually under 10 minutes.

No similar episodes found.

Frequently Asked Questions

How long is this episode of AI可可AI生活?

This episode is 27 minutes long.

When was this AI可可AI生活 episode published?

This episode was published on October 31, 2025.

Can I download this AI可可AI生活 episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.
URL copied to clipboard!