Training models to leverage CoT for improved performance on benchmark datasets.
Boost LLM reasoning with Chain-of-Thought Prompting.