Все публикации

0:27:38

Steering vectors: tailor LLMs without training. Part I: Theory (Interpretability Series)

0:33:00

Steering vectors: tailor LLMs without training. Part II: Code (Interpretability Series)

0:21:02

Decoding hidden states of Phi-3 with LogitLens (Interpretability Series)

0:38:11

State Space Models (S4, S5, S6/Mamba) Explained

0:32:12

Influence functions for large language models - why LLMs generate what they generate

0:32:06

Three times artificial neural networks are nothing like the human brain (+ are they ever alike?)

0:34:36

Does ChatGPT memorize train data? - exploring memorization in neural networks

0:22:55

Bounding the generalisation error in machine learning with concentration inequalities

0:40:54

A very, very basic coding tutorial for distributed optimization

0:21:17

A very, very basic introduction into distributed optimization

0:57:27

Efficient distributed optimization with mirror descent + a mirror descent introduction

0:02:54

To interact or not? The convergence properties of interacting stochastic mirror descent.