Large Language Models or LLMs often require guidance when dealing with tasks that call for multifaceted, multi-step reasoning. This factor holds the key when interacting with complex requests. A recent study introduces Step-Back Prompting, a new prompt engineering technique that promises to improve the correctness of intermediate reasoning steps through a process of supervision, which is conducted step-by-step. Step-Back Prompting is aimed at enhancing the model's reasoning capabilities by allowing it to revisit its past steps, verifying the correctness before proceeding further. This approach significantly contrasts with the well-known chain-of-thought reasoning, where decisions are made in a more linear, sequence-based manner. An extensive comparison between Step-Back Prompting and Chain-Of-Thought is detailed in this study. Through examples, the study encapsulates the benefits and methodology of Step-Back Prompting, demonstrating how it thrives on complex queries that require a deep understanding of subject matter or principles. The early results showcase a substantial improvement that this technique offers, making it a promising addition to the LLM field.