A Mechanistic Interpretation of Syllogistic Reasoning in Auto-Regressive Language Models

Open in new window