ViewTube

What can we learn about an AI by reading its reasoning—and what stays hidden inside the model? For Episode 5 of AI Deep Dive ...

Language models are increasingly capable of complex reasoning, yet their successes and failures offer only a partial picture of ...

Goodfire researcher Eric Bigelow joins the show to investigate how large language models arrive at decisions at the level of ...

Mechanistic interpretability explained: How do LLMs think? Wish we knew. But no, for real. My unpopular take: you can't hate AI ...

Machine learning. Data Science. KNIME tools for interpreting predictive machine learning models. This tutorial focuses on two ...

Interpretable Deep Learning Seminar Series Speaker: Chudi Zhong (University of North Carolina) Title: Breaking the One Model ...

The global-workspace finding quietly inverts interpretability's own premise: instead of us decoding the model's internals, it points ...

Recorded live at COLM 2026 in San Francisco. Siyan Li, a fourth-year PhD student at Columbia University, joins Jekaterina ...

Anthropic trained sparse autoencoders on Claude 3 Sonnet and pulled out millions of interpretable features — up to 34 million.

Host: Dr. Mengjia Xu (Assistant Professor, Department of Data Science, NJIT) Abstract: Time series data are central to high-stakes ...

Hand an AI investigator advanced interpretability tools and it can still fail, because it cannot separate the signal from the noise.

Why can the same factor have different associations across places and years? The Coefficient That Moves follows the journey ...

Discover how understanding advanced AI language models could unlock secrets of the human brain. This research explores the ...

Get the full NCP-AAI course on Udemy (with certificate of completion): ...

AI in Health: Foundation Models for Radiology, Cardiology and Autism Diagnosis This talk explores how AI can transform medical ...

AudioLens is an open-source workbench for understanding speech models. Load a clip and see where the model listened, what it ...

What can one average coefficient hide? From global to space-time in 46 seconds: • Global: one reference coefficient across ...

Support the Channel - Buy Me a Coffee: https://buymeacoffee.com/jod35 - Patreon: https://www.patreon.com/jod35 - Become a ...