2 papers
cs.SD2026
DeSRPA: Decoupled Speech Role-Playing Agent via Inference-Time Intervention
Wenqiu Tang, Zhen Wan, Takahiro Komamizu +1
While Large Language Models (LLMs) have revolutionized text-based role-playing, creating immersive Speech Role-Playing Agents (SRPAs) requires a seamless bridge between cognitive r…
cs.CL2026
Facet-Level Persona Control by Trait-Activated Routing with Contrastive SAE for Role-Playing LLMs
Wenqiu Tang, Zhen Wan, Takahiro Komamizu +1
Personality control in Role-Playing Agents (RPAs) is commonly achieved via training-free methods that inject persona descriptions and memory through prompts or retrieval-augmented…