Signal2026-07-21
arXiv

Explaining and Tuning Transformer-based LLMs in Arithmetic Tasks with Human Strategies

Open primary sourceRead on Substack