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

Part of

LLM Self-Distillation For Enhanced Reasoning