Open-source EU AI Act Annex IV documentation toolkit. Mechanistic interpretability + circuit discovery for transformers. One function call generates a structured, hash-chained evidence package.
mcp pytorch alignment sparse-autoencoders sae black-box-testing explainability fastapi gpt2 regulatory-compliance mechanistic-interpretability transformerlens eu-ai-act compliance-audit llm-compliance transformer-circuits logit-lens attribution-patching circuit-discovery annex-iv
-
Updated
Jun 15, 2026 - Python