Tag
This paper introduces a method for LLM-based chatbots to dynamically generate in-group personas by first identifying a user's primary concern and then creating a synthetic persona that shares that concern. A human-subject study demonstrates significant improvements in perceived rapport and user engagement compared to baseline conditions.
Introduces Persona Policies (PPol), a plug-and-play control layer that uses LLM-driven evolutionary program search to generate diverse, human-like user personas for evaluating LLM agents. Achieves 33–62% fitness gains over baseline, with human-likeness rated at 80.4%, and improves agent robustness with +17% task success.
NVIDIA's Nemotron-Personas-Korea is a dataset of 6-7 million synthetic personas grounded in official Korean demographic statistics, designed to help build culturally accurate Korean AI agents while complying with Korea's Personal Information Protection Act (PIPA). The tutorial demonstrates how to filter personas and deploy a grounded Korean AI agent using hosted APIs in approximately 20 minutes.