Multi-party Goal Tracking with LLMs: Comparing Pre-training, Fine-tuning, and Prompt Engineering

A dataset of 29 multi-party conversations between patients, their companions, and a social robot in a hospital.

BibTex: