Asking the model to think through a problem before answering improves accuracy on math, logic, and multi-step tasks.
25
Intermediate
Ct
Chain of Thought
Prompting
Chain of Thought
Having the model reason step by step.
Having the model reason step by step.
Asking the model to think through a problem before answering improves accuracy on math, logic, and multi-step tasks.