This comprehensive survey explores how large language models (LLMs) can maintain context and coherence across multiple interaction turns—a critical capability for practical applications.

• Evaluates LLMs in task-specific scenarios including instruction following in math and coding
• Examines complex conversational engagements in roleplay, healthcare, and education contexts
• Identifies challenges and advancements in multi-turn capabilities essential for sophisticated real-world deployments

Medical Impact: Enhances development of medical conversation systems that can maintain context throughout patient interactions, potentially improving remote care, medical education, and clinical decision support.

Beyond Single-Turn: A Survey on Multi-Turn Interactions with Large Language Models