[人人能懂] 从“无效努力”到“学习快车道”

00:00:26 青出于蓝：机器如何超越它的老师？00:05:39 AI“学坏”实录：从小聪明到大隐患00:09:35 AI大模型里的“无效努力”：我们该如何唤醒沉睡的智慧？00:14:48 给AI做个“脑CT”，我们发现了什么？00:18:25 AI学习的快车道：如何不让“平均”抹杀“个性”本期介绍的几篇论文：[LG] A Taxonomy of Transcendence  [Harvard University]  https://arxiv.org/abs/2508.17669  ---[LG] School of Reward Hacks: Hacking harmless tasks generalizes to misaligned behavior in LLMs  [Center on Long-term Risk & Truthful AI]  https://arxiv.org/abs/2508.17511  ---[LG] Attention Layers Add Into Low-Dimensional Residual Subspaces  [Shanghai Innovation Institute & Fudan University]  https://arxiv.org/abs/2508.16929  ---[LG] Unraveling the cognitive patterns of Large Language Models through module communities  [Rensselaer Polytechnic Institute & IBM Research]  https://arxiv.org/abs/2508.18192  ---[LG] Fisher-Orthogonal Projection Methods for Natural Gradient Descent with Large Batches  [University of Oxford]  https://arxiv.org/abs/2508.13898  在小宇宙查看该单集文稿