works on

From the 1 of 4 linked papers with an AI index.

collaborators

6 papers

cs.CL2026

What Models Express, Suppress, and Resist: Auditing Open-Weight LLMs with Persona Vectors

Winston Zeng, Ali Emami, Jinho D. Choi

The paper introduces a large inventory of persona vectors to systematically probe open-weight language models, categorizing traits as naturally expressed, steerable, or resistant,…

cs.AI2026

Hierarchical Multi-Persona Induction from User Behavioral Logs: Learning Evidence-Grounded and Truthful Personas

Nayoung Choi, Haeyu Jeong, Changbong Kim +2

Behavioral logs provide rich signals for user modeling, but are noisy and interleaved across diverse intents. Recent work uses LLMs to generate interpretable natural-language perso…

cs.HC2026

Designing and Evaluating a Conversational Agent for Early Diagnosis of Alzheimer's Disease and Related Dementias

Andrew G. Breithaupt, Nayoung Choi, James D. Finch +5

Early diagnosis of Alzheimer's disease and related dementias (ADRD) is critical for timely intervention, yet most diagnoses are delayed until advanced stages. While comprehensive p…

cs.CL2025

Generative Induction of Dialogue Task Schemas with Streaming Refinement and Simulated Interactions

James D. Finch, Yasasvi Josyula, Jinho D. Choi

In task-oriented dialogue (TOD) systems, Slot Schema Induction (SSI) is essential for automatically identifying key information slots from dialogue data without manual intervention…

cs.CL2024

Transforming Slot Schema Induction with Generative Dialogue State Inference

James D. Finch, Boxin Zhao, Jinho D. Choi

The challenge of defining a slot schema to represent the state of a task-oriented dialogue system is addressed by Slot Schema Induction (SSI), which aims to automatically induce sl…

cs.CL2024

Diverse and Effective Synthetic Data Generation for Adaptable Zero-Shot Dialogue State Tracking

James D. Finch, Jinho D. Choi

We demonstrate substantial performance gains in zero-shot dialogue state tracking (DST) by enhancing training data diversity through synthetic data generation. Existing DST dataset…