Tag
The paper develops a framework for prompt design in LLM-based robots to prevent hallucinated capabilities and misleading personas, introducing a structured template with eight functional components based on expert surveys and literature review.