Part IV: Training & Adapting LLMs
As large language models are deployed in high-stakes applications, the question "why did the model produce this output?" becomes critical. Interpretability research aims to open the black box of transformer models, revealing the internal computations that drive predictions, the features that neurons encode, and the circuits that implement specific behaviors.
This module covers the full spectrum of interpretability methods for transformers. It begins with attention analysis and probing classifiers, which offer accessible entry points for understanding model internals. It then advances to mechanistic interpretability, the ambitious program of reverse-engineering neural networks at the level of individual features and circuits. The module also covers practical interpretability tools for debugging, model editing, and representation engineering, as well as formal attribution methods for explaining transformer predictions.
By the end of this module, you will be able to analyze attention patterns to understand model behavior, use probing classifiers to test what information is encoded in hidden states, apply sparse autoencoders to extract interpretable features, and employ attribution methods to explain individual predictions.