A prompting technique where the model is asked to reason step-by-step before answering, often producing dramatically better results on math/logic tasks.
A prompting technique where the model is asked to reason step-by-step before answering, often producing dramatically better results on math/logic tasks.