KnowledgeDebugger -- an Exploration Tool for Knowledge Localization and Editing in Transformers
2026-07-01 • Computation and Language
Computation and Language
AI summaryⓘ
The authors created KnowledgeDebugger, a simple tool with a visual interface to help people explore and change what Transformers know. This tool makes it easy to try out different ways of editing knowledge without needing to write code. They show that their tool works well by testing it on examples from recent research. The goal is to help researchers better understand and experiment with how Transformers store and update information.
TransformerKnowledge EditingKnowledge LocalizationGUILM-DebuggerEasyEditState-of-the-artMachine LearningNatural Language ProcessingModel Interpretability
Authors
Eric Benz, Lennart Stöpler, Nikolai Bolik, Artur Andrzejak
Abstract
Recent research has increasingly focused on understanding how Transformers store and process knowledge, as well as how this knowledge can be edited. Research work in this area is often conducted in two phases: first, phenomena are explored on individual samples. Then, when results appear promising, more statistically robust experiments follow. To support the first phase, we propose KnowledgeDebugger, a GUI-based exploration tool for knowledge localization and editing in Transformers. Our tool - inspired by LM-Debugger - offers no-code access to the methods in EasyEdit, a widely used library of state-of-the-art Knowledge Editing approaches. We demonstrate the tool's effectiveness through case studies of recent findings in this field.