Fetching the paper…

Steering Large Language Models between Code Execution and Textual Reasoning · Around