This is largely due to the fact that current LLMs often struggle with complex code, multi-step logic, and abstract tasks, frequently exhibiting logical leaps, disorganized steps, and irrelevant ...