New framework AdaMARP boosts LLM role-playing immersion

Researchers have developed AdaMARP, a novel adaptive multi-agent interaction framework designed to enhance immersion and adaptability in large language model (LLM) role-playing scenarios. The framework introduces an immersive message format and a Scene Manager to govern role-playing actions, scene transitions, and character introductions. Experiments show that AdaMARP, particularly with an 8B actor model, improves character consistency and narrative coherence, and a 14B LLM utilizing AdaSMSet outperforms Claude Sonnet 4.5 in orchestrating role-playing interactions. AI

IMPACT Enhances immersion and adaptability in LLM role-playing, potentially improving interactive narrative experiences.

RANK_REASON The cluster contains a research paper detailing a new framework for LLM role-playing. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• AdaMARP
• arXiv
• Claude Sonnet 4.5
• Hugging Face
• Zhenhua Xu

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →