Extracting Interpretable Task-Specific Circuits from Large Language Models for Faster Inference

Re-assign community