Implementing Transparent AI: Model Explainability

Lightning Lessons

Practical skills & tools to accelerate your career

Implementing Transparent AI: Model Explainability

Hosted by Bhaskarjit Sarmah

What you'll learn

Decoding Neuron Functions

Use activation probing and feature visualization to link neurons to human-understandable concepts.

Tracing Neural Circuits

Analyze weight matrices and activations to reconstruct neuron chains driving specific model behaviors.

Linking Circuits to Model Behavior

Correlate circuit findings with model outputs to diagnose errors and suggest targeted interventions.

Why this topic matters

Mechanistic interpretability demystifies how AI makes decisions, boosting trust and accountability. Professionals who master it can debug complex models faster, ensure fairness, and meet regulatory standards. This lesson helps you unlock deeper model insights, enhancing your expertise and career potential.

You'll learn from

Bhaskarjit Sarmah

I’m a Director at BlackRock with 10+ yrs in AI/ML & awarded Top 5 GenAI Leader.

Add more about your accomplishments, work history, and credentials, especially those that help demonstrate your credibility as an instructor. Consider including details about why you want to share your expertise and how students will get value from learning with you.

Previously at BlackRock