bycloud
bycloud · 269 个带字幕的视频bycloud 在 WinkLingo 收录了 269 个视频(这个频道有 24万 个订阅),每个视频都能逐词检索。
查看 YouTube 原频道英文简介
Frontier AI Research Explained
收录的视频
14:56LLM 中的 RL 讽刺(以及其疯狂的新元趋势)
The RL Irony in LLMs (and its insane new meta)
13:47AI 开源新三强
The New AI Open Source Trifecta
12:53DeepSeek V3.2 再次打破 SoTA... 它是怎么做到的?
DeepSeek V3.2 Just Broke SoTA Again… But How?
13:16一个 27M 的模型怎么可能击败 ChatGPT?
How did a 27M Model even beat ChatGPT?
10:35LLM(大语言模型)最大的谜团终于被解开了
The biggest Mystery of LLMs have just been solved
27:07中国 AI 的冰山之底
The Chinese AI Iceberg
8:27研究人员在训练大语言模型 (LLM) 方面变得非常有创意 [Token 顺序预测]
Researchers Are Getting Really Creative Training LLMs [Token Order Prediction]
12:27高效部署 LLM 的艺术:上下文工程详解
The Art of Serving LLMs Efficiently: Context Engineering Explained
11:16反向工程 AI 的训练数据真的可行吗?
Is It EVEN Possible To Reverse Engineer AI’s Training Data?
9:43AI新范式?!详解基于能量的Transformer (Energy-Based Transformers)
New AI Paradigm?! Energy-Based Transformers Explained
11:47AI 新趋势:训练 LLM 探索“困难”Token [RLVR + 熵]
New AI Meta: Train LLMs To Explore On "Hard" Tokens [RLVR + Entropy]
11:15Kimi K2 技术详解:它是如何挑战 AI 7 年现状的
Kimi K2 Technical Breakdown: How It Challenged AI’s 7-Year Status Quo
7:58如果 AI 这么强,为什么不能自己生成数据?[SEAL]
If AI's So Good, Why Can't It Make Its Own Data? [SEAL]
13:33比标准LLM快10倍!?DiffusionLM 详解
10x Faster Than Standard LLM!? DiffusionLM Explained
10:02用 ChatGPT 写作业过度会带来哪些副作用?
The Side Effects of Overusing ChatGPT For Homework
10:33POV:中国AI实验室教大家如何节省数百万美元
POV: Chinese AI Lab Teaching Everyone How To Save Millions of Dollars
9:34如果我们在大语言模型(LLM)中去掉分词(Tokenization)会怎样?
What If We Remove Tokenization In LLMs?
9:58当 AI 像生物大脑一样被设计时
When AI Is Designed Like A Biological Brain
15:34LLM 中我们未曾预料到的 RL 启示
The LLM's RL Revelation We Didn't See Coming
14:351-Bit LLM:最有效率的大语言模型?
1-Bit LLM: The Most Efficient LLM Possible?
10:13统一视觉与语言的真正AI架构
The REAL AI Architecture That Unifies Vision & Language
11:02DeepSeek 如何打造目前的“最强”数学证明 AI
How DeepSeek Built The Current "Best" Math Prover AI
2:30AI 新闻:DeepSeek-R1 V2,新的开源 SOTA!
AI News: DeepSeek-R1 V2, the new open source SoTA!
11:03为什么让 LLM 一次预测 4 个 token?详解多 Token 预测 (Multi-Token Prediction)
Why would anyone let LLMs predict 4 tokens at once? Multi-Token Prediction Explained