Random Chain-Of-Thought For LLMs & Distilling Self-Evaluation Capability