Yannic Kilcher Courses

On the Biology of a Large Language Model - Part 2

Delve into the internal mechanisms of Claude 3.5 Haiku as Yannic Kilcher explores Anthropic's Transformer Circuit Blog Post, examining how this lightweight production model functions across various contexts.

• YouTube
• 56 minutes
• Self-Paced
• Free Video

Byte Latent Transformer - Dynamic Patches vs Traditional Tokenization in Language Models

Delve into Meta's innovative LLM architecture that replaces traditional tokenization with dynamic byte-level patches, offering improved scaling and inference efficiency for language models.

• YouTube
• 36 minutes
• Self-Paced
• Free Video

Scaling Transformer to 1M Tokens and Beyond with RMT - Paper Explained

Explore the Recurrent Memory Transformer technique for scaling transformers to process 1 million tokens, examining its strengths, weaknesses, and potential impact on long-term dependency handling in NLP tasks.

• YouTube
• 25 minutes
• Self-Paced
• Free Video

OpenAssistant - ChatGPT's Open Alternative

Explore OpenAssistant, an open-source alternative to ChatGPT. Learn about the project's goals, data collection process, and participate in various AI interaction tasks to improve the system.

• YouTube
• 36 minutes
• Self-Paced
• Free Video

CICERO- An AI Agent That Negotiates, Persuades, and Cooperates With People

Explore Meta AI's Cicero, an advanced AI agent that excels in Diplomacy by integrating language models with strategic planning to negotiate and cooperate with humans effectively.

• YouTube
• 1 hour 1 minute
• Self-Paced
• Free Video

The New AI Model Licenses Have a Legal Loophole - OpenRAIL-M of BLOOM, Stable Diffusion, etc.

Análisis crítico de las licencias de IA responsable, sus limitaciones y posibles consecuencias legales. Explora el dilema entre el uso abierto y las restricciones, proponiendo una alternativa llamada OpenRAIL++.

• YouTube
• 28 minutes
• Self-Paced
• Free Video

ROME - Locating and Editing Factual Associations in GPT - Paper Explained & Author Interview

Explores mechanisms for storing and editing factual knowledge in GPT models, revealing key roles of mid-layer feed-forward modules and proposing targeted editing techniques with implications for model understanding and control.

• YouTube
• 1 hour 5 minutes
• Self-Paced
• Free Video

Is Stability Turning into OpenAI?

Explore the controversy surrounding Stability AI's actions, including community takeovers, intellectual property disputes, and concerns about openness in AI development.

• YouTube
• 39 minutes
• Self-Paced
• Free Video

Neural Networks Are Decision Trees - With Alexander Mattick

Exploring the equivalence between neural networks and decision trees, discussing interpretability, computational advantages, and potential collaborations between the two approaches.

• YouTube
• 32 minutes
• Self-Paced
• Free Video

AlphaTensor by DeepMind Explained

Explore DeepMind's AlphaTensor, a revolutionary AI system that discovers efficient matrix multiplication algorithms, outperforming human-designed methods and advancing computational efficiency across various fields.

• YouTube
• 55 minutes
• Self-Paced
• Free Video

How to Make Your CPU as Fast as a GPU - Advances in Sparsity with Nir Shavit

Explore advances in sparsity and how CPUs can match GPU performance in neural networks. Learn about pruning, efficient algorithms, and the future of sparse architectures in deep learning.

• YouTube
• 50 minutes
• Self-Paced
• Free Video

More Is Different for AI - Scaling Up, Emergence, and Paperclip Maximizers

Explore AI scaling, emergence, and safety with Jacob Steinhardt. Discuss qualitative changes in AI systems, implications for alignment, and thought experiments like the Paperclip Maximizer. Gain insights on future AI challenges and potential solutions.

• YouTube
• 1 hour 7 minutes
• Self-Paced
• Free Video

The Future of AI Is Self-Organizing and Self-Assembling - With Prof. Sebastian Risi

Explore self-organizing AI systems, swarm intelligence, and neural cellular automata with Prof. Sebastian Risi. Discover applications, challenges, and future directions in this emerging field of artificial intelligence.

• YouTube
• 1 hour 2 minutes
• Self-Paced
• Free Video

A Path Towards Autonomous Machine Intelligence - Paper Explained

Explanation of Yann LeCun's paper on autonomous machine intelligence, combining self-supervised learning, energy-based models, and hierarchical predictive embedding for creating world models and planning.

• YouTube
• 1 hour
• Self-Paced
• Free Video

Learning to Act by Watching Unlabeled Online Videos - Paper Explained

Exploring OpenAI's Video PreTraining for Minecraft: leveraging unlabeled gameplay footage to create an AI agent capable of complex tasks, including crafting a diamond pickaxe.

• YouTube
• 33 minutes
• Self-Paced
• Free Video