2025/11/15 07:36OpenAI Researchers Train Weight Sparse Transformers to Expose Interpretable Circuitsvia MarkTechPost (author: Michal Sutter)