This article explores visual debugging tools for machine learning workflows. It discusses what to visualize during training, available visualization tools, and methods to capture model computations using hooks and breakpoints.
This research paper addresses the limitation of users interacting with language models via single outputs, which hides the full distribution of possible generations. It introduces GROVE, an interactive visualization that represents multiple LM generations as overlapping paths in a text graph, revealing shared structure and branching points for better understanding.
This content addresses the challenge of understanding modern AI model architectures, particularly within Hugging Face repositories. It introduces a method to instantly visualize these architectures, aiming to simplify the comprehension of internal model structures for developers.