DeSRPA introduces a dual-level control vector method for inference-time intervention on frozen backbones to improve personality consistency and speech naturalness in role-playing agents over end-to-end fine-tuned baselines.
Facet-level persona control by trait-activated routing with contrastive sae for role- playing llms,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
DeSRPA: Decoupled Speech Role-Playing Agent via Inference-Time Intervention
DeSRPA introduces a dual-level control vector method for inference-time intervention on frozen backbones to improve personality consistency and speech naturalness in role-playing agents over end-to-end fine-tuned baselines.