Version updated for https://github.com/designer-coderajay/Glassbox-AI-2.0-Mechanistic-Interpretability-tool to version v4.2.2. This action is used across all versions by ? repositories. Action Type This is a Composite action. Go to the GitHub Marketplace to find the latest changes. Action Summary Glassbox is an open-source toolkit designed to automate the creation of technical documentation required for compliance with the EU AI Act Annex IV, streamlining regulatory processes for high-risk AI systems. It also offers advanced tools for researchers to analyze transformer models, enabling efficient discovery of minimal faithful circuits responsible for predictions. This action supports both white-box and black-box AI models while providing structured outputs to aid compliance and interpretability efforts.

Version updated for https://github.com/designer-coderajay/Glassbox-AI-2.0-Mechanistic-Interpretability-tool to version v4.2.1. This action is used across all versions by ? repositories. Action Type This is a Composite action. Go to the GitHub Marketplace to find the latest changes. Action Summary Glassbox is an open-source toolkit designed to automate the creation of EU AI Act Annex IV compliance documentation for high-risk AI systems, compatible with both open-source models and proprietary APIs like GPT-4. It simplifies regulatory compliance by generating structured, nine-section technical documentation drafts, while also offering researchers tools to analyze and identify minimal faithful circuits in transformer models for improved interpretability. Key capabilities include multi-framework support, cross-model comparison, and significantly faster performance for mechanistic interpretability tasks.