What Matters Right Now In Mechanistic Interpretability

Summary

This is a talk I gave to my MATS 9.0 training scholars about the big picture of mech interp - as of Oct 2025, what had changed? A practical guide to neural nets, superposition, and expts around polysemanticity. Try Explain by Scrimba: ... Explaining "Neural Networks are Already Clean" from the ground up. Cline Desktop: cline.gg/desktop-avb and ... Take your personal data back with Incogni! Use code WELCHLABS at the link below and get 60% off an annual plan: ... What if you were to peer inside the 'mind' of AI? You wouldn't find fully formed thoughts, My book about Post-Labor Economics - "Labor Zero" is A discussion on the philosophy of deep learning, 80000hours.org/mlst Visit our sponsor 80000 hours - grab their free career guide and their podcast! Use our ... The Cohere For AI community was honoured to welcome Catherine Olsson to discuss the process of getting started in SITUATION EXPLAINED: Ex-OpenAI researcher Nick Cammarata says we can How can we look inside neural networks and figure out how they do what they do? This is likely to be very important for alignment ... How can we use the language of causality to understand and edit the internal mechanisms of AI models? Atticus Geiger ... CS 7180: Neural Mechanics Spring 2026 Course at Northeastern University Modern AI systems are powerful but opaque: even ... Neel Nanda from DeepMind presenting '